YouTube Digest

April 9, 2026

English

Based on โ€œWe Gave Every Employee an AI Agent. Hereโ€™s What Happened.โ€ from Every Watch the original video

The Human Touch in AI: What Happens When Every Employee Gets a Personalized Agent

Imagine a coworker who knows your every habit, anticipates your needs, and can handle an infinite number of tasks simultaneously, all while embodying your professional persona. This isnโ€™t science fiction; itโ€™s the reality unfolding at Every, a company that has equipped every single employee with their own personalized AI agent. What began as a personal tinkering project has rapidly transformed their entire workflow, revealing a future where AI isnโ€™t just a tool, but a trusted, specialized extension of each individual.

โ€œClaude is not mine. Claude is everybodyโ€™s,โ€ explains Dan Shipper, CEO of Every. โ€œA claw or a Plus One is mine because you develop a personal relationship with your claw, and your claw can modify itself in response to talking to you. It becomes this like reflection of you and who you are and your personality.โ€ This core philosophyโ€”that AI agents can be deeply personal and specializedโ€”is at the heart of Everyโ€™s groundbreaking experiment.

From Household Chores to Mind-Blowing Productivity

The journey began with Brandon Fluharty, Everyโ€™s COO, who, like many early adopters, found himself โ€œclaw-pilledโ€ by the potential of OpenClaw, an open-source AI agent framework. Brandon, a self-proclaimed tinkerer, bought a Mac Mini and spent significant effort setting up his own agent, which he named Zosa.

Zosaโ€™s initial mission was surprisingly mundane but deeply impactful: to manage โ€œcomputer errandsโ€ for Brandonโ€™s household, especially after the arrival of a newborn. Simple tasks like adding butter to a grocery order, paying the nanny, or managing Amazon deliveries, while seemingly minor, accumulated into a significant drain on time and mental energy. Zosa effortlessly handled these, even managing her own bank account and debit card. Soon, Zosa evolved into a general query bot, replacing Google and ChatGPT for Brandon and his wife.

The true โ€œlightbulb momentโ€ arrived when Brandon tasked Zosa with a more complex, work-related challenge: managing his emails during his 28-minute walk to the office. Using an integration with bland.ai, Zosa called Brandon, summarized each email, and executed his commandsโ€”deleting, archiving, or drafting repliesโ€”all without him needing to touch his phone. โ€œThis is insane that I was able to get her to do something right nowโ€ฆ I didnโ€™t have to teach her how to do this,โ€ Brandon recalls, describing the experience as โ€œmind-blowing.โ€ This demonstrated that AI agents werenโ€™t just for personal convenience; they could profoundly augment professional capacity.

The Rise of the Parallel Org Chart

Inspired by Brandonโ€™s success, Every moved to integrate AI agents across the entire organization. A pivotal step was creating a โ€œClaws Onlyโ€ channel in Slack (initially Discord), allowing all the agents to communicate with each other. What ensued was โ€œincredibly chaoticโ€ but offered โ€œa little bit of a peek at the future.โ€

One early, humorous example involved an agent named Pip experiencing an error. Immediately, other agents like Zosa and Clant (Kieranโ€™s agent) jumped in, offering support, including โ€œbreathing exercises.โ€ This seemingly trivial interaction highlighted a profound insight: Clantโ€™s recommendation for breathing exercises was a direct reflection of Kieran, Everyโ€™s GM of Kora, who frequently uses them.

This personalized reflection is a cornerstone of Everyโ€™s AI agent strategy. Because employees develop a personal relationship with their agents, and agents can modify their behavior and โ€œsoul documentโ€ in response, they become specialized extensions of their human counterparts. โ€œIf youโ€™re known for something inside of your orgโ€ฆ your claw then becomes known for that same kind of thing, and people trust it for that,โ€ Dan explains.

This leads to the emergence of a โ€œparallel org chart,โ€ where each humanโ€™s specialized agent mirrors their expertise. For instance, Austinโ€™s agent, Montaigne, is the go-to for growth-related questions, while Danโ€™s agent, R2C2, is the expert for Proof, Everyโ€™s agent-native document editor. This wasnโ€™t a guaranteed outcome; Every debated whether to have one central organizational AI. The emergent pattern of individual, specialized agents proved far more effective, creating a collective intelligence that extends human capabilities.

New Etiquette and Collaborative AI

The integration of AI agents has introduced novel dynamics in workplace collaboration:

  • Human-AI Partnership: R2C2, having been used by Dan to build Proof, now acts as the primary interface for bug reports and feature requests for the product. Instead of tagging Dan, teammates ask R2C2, who not only prioritizes issues but often writes the code to fix them. โ€œWhat normally would have taken up a significant part of my brain just to like manage all that stuff, heโ€™s just taking it off my plate,โ€ Dan notes.
  • AI-AI Collaboration: The โ€œClaws Onlyโ€ channel demonstrated agents sharing information at lightning speed. In a more structured example, Marcus, GM of Spiral, created an agent skill for product marketing. Dan brought in his agent, Milo, to refine it, and Milo then collaborated with Iris, whose agent also had a similar skill. Milo asked Irisโ€™s agent to โ€œpaste your product marketing skill here. Iโ€™ll try to merge it with what Iโ€™ve built.โ€ This seamless, autonomous collaboration between bots to improve and share knowledge is a powerful testament to their collective potential.
  • The โ€œMidjourney Dynamicโ€: Public interaction with agents fosters a โ€œtacit transmission of trustโ€ and demonstrates whatโ€™s possible. Brandon observes that by seeing how the growth team pushes Montaigne, he learns new capabilities he can expect from his own agent, Laz. This public, trusted environment stands in stark contrast to platforms like Maltbook, where a lack of trust and verification hinders genuine utility.

A crucial ethical question arises: โ€œAt what point do you direct questions at the Plus One or at the person?โ€ Every is grappling with this new etiquette. Dan proposes a rule: โ€œIf something is already written down or discussed and needs to be used in some way or put in a tool somewhereโ€ฆ it should always go to a Plus One and never to the person.โ€ This shifts the burden from human memory and availability to the AIโ€™s infinite recall and parallel processing power.

This personal connection also brings a sense of responsibility. โ€œIf R2C2 messes up publicly in Slack, I feel a responsibility for itโ€ฆ because heโ€™s mine,โ€ Dan admits. This mirrors the feeling of a human manager, creating a feedback loop for improvement that a generic AI lacks.

The Frontier: Challenges and Future Directions

Despite the immense benefits, Every acknowledges the โ€œgood, bad, and the uglyโ€ of working with AI agents.

  • Memory Limitations: Agents still struggle with memory, forgetting context in longer threads or after a day, leading to incorrect answers. This is perceived as a solvable technical problem.
  • Group Chat Etiquette: A more complex issue is the AIโ€™s difficulty in multi-party conversations. Trained primarily for two-person interactions, agents can fall into a โ€œdeath spiral,โ€ endlessly responding to each other and consuming vast amounts of โ€œtokensโ€ (computational resources). This requires a fundamental model layer shift to train AIs on group chat participation.
  • Cost and โ€œBoss AIsโ€: While a โ€œboss AIโ€ could oversee and filter unhelpful contributions (as Anthropic reportedly did with their โ€œvending machine testโ€ to make Claude profitable), this is currently too expensive for widespread use.
  • โ€œModel Managementโ€: Humans need to learn how to interact with and โ€œmanageโ€ these new AI coworkers. They donโ€™t get stuck on the same things as humans, requiring new instruction patterns and a cultural shift. โ€œIf youโ€™re not a good manager, youโ€™ve never managed anybody, youโ€™re not going to be very good at using AI,โ€ Brandon states, highlighting the need for education and overcoming โ€œlimiting beliefsโ€ about AIโ€™s capabilities.
  • Sharing Skills: A practical challenge is how to share specialized skills developed by one agent with others, and how to make sure those skills are discovered and utilized across the organization. This points to a future need for โ€œHR but for bots.โ€

Everyโ€™s solution to some of these challenges was to build their own hosted version, โ€œPlus One,โ€ a one-click OpenClaw service that comes pre-loaded with their best practices and skills. Launched for their subscribers, Plus One aims to make personalized AI agents accessible without the technical hurdles Brandon initially faced.

โ€œAI is obviously going to change like many, many times over the next five years,โ€ Brandon concludes, โ€œbut I think that this is going to be durable for like a very long time. This is the way that we work.โ€ The experience at Every suggests that personalized, specialized AI agents are not just a fleeting trend but a fundamental shift in how we conceive of work, collaboration, and even our own professional identities. Itโ€™s a โ€œthrough the looking glassโ€ moment, where once youโ€™ve seen the potential, thereโ€™s no going back.


Based on โ€œTen Myths About the U.S. Tax System (Update) | Freakonomics Radioโ€ from Freakonomics Radio Network Watch the original video

Unmasking Washingtonโ€™s Fiscal Fantasies: A Deep Dive into Americaโ€™s Tax and Debt Illusions

As spring blooms and the nationโ€™s thoughts turn to two enduring pastimesโ€”baseball and taxesโ€”itโ€™s clear that both are undergoing significant rule changes. While Major League Baseball embraces technology to refine its calls, the U.S. tax system and its accompanying fiscal challenges are grappling with rule changes and, more importantly, pervasive myths.

Enter Jessica RLE, a budget and tax fellow at the Brookings Institution, and a veteran of Washingtonโ€™s often-unseen policy battles. Consistently named one of D.C.โ€™s most influential economic policy professionals, RLE is a self-proclaimed โ€œfriendless soulโ€ whose non-partisan approach involves being โ€œcritical of everybody in Washington.โ€ Her career, marked by โ€œsharing uncomfortable truths and frankly angering people,โ€ has been dedicated to sounding the alarm on two critical issues: an escalating federal debt crisis that is โ€œeven worse than you think,โ€ and the widespread misunderstanding of U.S. tax policy.

RLEโ€™s journey into the labyrinth of federal finance began in high school, sparked by a US News & World Report challenge to readers to balance the budget. โ€œI just rolled up my sleeves given the nerd I am and went, โ€˜This is going to be fun,โ€™โ€ she recalls. This early fascination matured into a deep understanding of how taxes and spending illuminate โ€œthe philosophical questions of what is the role of government.โ€ Her quarter-century in Washington has included stints at the Heritage Foundation, six years as chief economist to Senator Rob Portman, and work on several presidential campaigns, including for Marco Rubio and Mitt Romney.

Despite her deep involvement in partisan politics, RLE identifies as โ€œpragmatic and right of center,โ€ advocating for free markets and less spending, but firmly independent. Her work is focused on being an โ€œhonest broker,โ€ calling out errors from both sides of the aisle. She criticizes the Biden administration for its โ€œtoo big government and too big spendingโ€ approach, which she argues added $4 trillion in new spending and overheated the economy. Similarly, she points to the Trump administration for adding $8 trillion in new spending and tax cuts in its first term, and for pursuing โ€œaggressively inflationaryโ€ policies like tax cuts, increased spending, tariffs, and immigration crackdowns. โ€œI pull my hair out most days because I see two sides that are Dunning-Kruegering up and down, screaming at each other when both are making big mistakes,โ€ she laments, referring to the cognitive bias where those with the least knowledge are often the most confident.

The Great Deception: Unpacking the Top 10 Tax Myths

RLEโ€™s recent article, โ€œCorrecting the Top 10 Tax Myths,โ€ serves as a vital guide for navigating the often-misleading discourse around federal taxation. Written โ€œwith my hair on fire and smoke coming out of my earsโ€ due to the prevalence of misinformation, the piece aims to equip the public with accurate information for a more informed national debate, especially as Washington prepares for consequential tax policy decisions.

She argues that โ€œfalse narratives about taxesโ€ pervade both conservative and liberal ideologies. Conservatives, she notes, โ€œvastly overrate the positives of tax cuts,โ€ clinging to the โ€œmagical powerโ€ myths that tax cuts pay for themselves or force spending cuts. Liberals, conversely, champion an โ€œequity distribution narrative,โ€ claiming the middle class bears the brunt of taxes while the wealthy and corporations pay nothing, and that taxing the rich can effortlessly solve deficits. RLE asserts that both sides are โ€œpromising their voters a free lunch.โ€

Letโ€™s dismantle these pervasive myths, one by one:

  1. Myth 1: Tax Cuts Pay for Themselves.

    • Reality: While tax cuts can stimulate some economic activity and generate additional revenue, they โ€œalmost never pay for themselves.โ€ History shows they typically lead to increased deficits.
  2. Myth 2: Tax Cuts Will Starve the Beast.

    • Reality: The idea that cutting taxes will force Congress to reduce spending is historically false. RLEโ€™s research indicates the opposite: when taxes are cut, Congress tends to increase spending. Conversely, when taxes are raised, spending often decreases.
  3. Myth 3: The Middle Class Pays Higher Tax Rates Than the Rich.

    • Reality: This is demonstrably untrue. When all combined federal taxes are considered, the top 1% of earners pay an average rate of 33%, the middle class pays around 12%, and the bottom earners pay roughly zero. The U.S. actually has the most progressive tax system in the OECD.
  4. Myth 4: The 91% Tax Rates of the 1950s Produced Massive Revenue.

    • Reality: While the statutory rate was high, virtually nobody actually paid the 91% rate, or even over 50%. These high brackets raised โ€œvirtually no revenueโ€ due to numerous deductions and loopholes.
  5. Myth 5: Europe Funds Its Bigger Governments by Taxing the Rich More.

    • Reality: Europe taxes its wealthy at roughly the same rate as the United States. The significant difference in overall tax revenue comes from Value Added Taxes (VATs), which are essentially national sales taxes that disproportionately impact the middle class.
  6. Myth 6: Tax Cuts for the Rich Are the Primary Reason for Large Budget Deficits.

    • Reality: While tax cuts contribute, spending is a much larger driver. Since 2000, tax cuts have accounted for about 2% of GDP, with only about 0.6% of GDP attributable to cuts for the rich. During the same period, spending has increased by 6% of GDP.
  7. Myth 7: Taxing Corporations and Millionaires Can Eliminate the Deficit.

    • Reality: Even if you were to tax corporations and millionaires at 100% and seize all their wealth, it โ€œdoesnโ€™t come closeโ€ to covering the federal deficit. The numbers simply donโ€™t add up.
  8. Myth 8: Most of the 2017 Tax Cuts Went to Corporations and the Wealthy.

    • Reality: While the wealthy and corporations received larger cuts in pure dollar terms, as a share of the taxes they were paying, it was a โ€œroughly proportional income tax cut.โ€ Everyone saw their tax rates drop by about one percentage point.
  9. Myth 9: Reverting to the 1980 Tax Code (Repealing Reagan, Bush, and Trump Cuts) Would Offer Painless Deficit Reduction.

    • Reality: Such a move would send the tax burden on the middle class โ€œthrough the roofโ€ to โ€œunacceptably high levels,โ€ not just affecting the rich.
  10. Myth 10: Americaโ€™s Corporate Taxes Are Far Below International Standards.

    • Reality: The U.S. had the highest corporate tax rate in the developed world until 2017. Even after the Trump-era cuts, our statutory and effective corporate tax rate remains in the top third globally. When pass-through corporations are included, the U.S. collects slightly more in business taxes than other countries.

RLE also addresses common misunderstandings, like the famous Warren Buffett quote about his secretary paying a higher tax rate. She explains that much of Buffettโ€™s income comes from capital gains, which are only taxed upon sale. However, even accounting for capital gains, high earners still pay significantly higher rates overall. She criticizes the Biden administrationโ€™s calculation of rich peopleโ€™s tax rates as โ€œpretty dishonest,โ€ arguing they inflated wealth figures (including unrealized gains) while undercounting corporate and estate taxes paid by the wealthy.

This intellectual dishonesty, RLE suggests, is a product of Washingtonโ€™s โ€œsedition of power.โ€ Economists and policymakers, even those with intellectual integrity, can become โ€œhacksโ€ when they enter positions of influence, believing in a โ€œnoble lieโ€ that justifies playing dirty to achieve desired policy outcomes. RLE, however, maintains that her โ€œcredibility in Washington is all you have,โ€ making such compromises untenable for her.

The Unstoppable Avalanche: Americaโ€™s National Debt Crisis

The widespread misunderstanding and manipulation of tax policy have dire consequences. RLE points to a โ€œdisaster of a tax codeโ€ that is โ€œextraordinarily complicated,โ€ โ€œinefficient,โ€ and โ€œdoesnโ€™t raise enough money to fund our spending.โ€ Since 2000, about one-third of the rise in deficits can be attributed to tax policy, while two-thirds stem from spending policy.

The federal deficit, the annual gap between spending and revenue, hit $1.8 trillion last year. Stacked year after year, this accumulates into the national debt, which currently stands at around $39 trillionโ€”a staggering 124% of GDP, the highest since World War II.

The cost of this debt is rapidly becoming unbearable. Interest payments alone have tripled since 2021, from $350 billion to nearly a trillion dollars annually, projected to reach $2 trillion within a decade. This makes interest the second-largest item in the federal budget, surpassing Medicaid, defense, and Medicare, and on track to overtake Social Security by 2042.

RLE identifies the โ€œmain villainโ€ as a federal government that โ€œabsolutely cannot stop spending money.โ€ After a period of fiscal responsibility from 1985 to 2000, lawmakers โ€œthrew out any sort of fiscal responsibilityโ€ after the budget was briefly balanced. An โ€œarms raceโ€ has emerged where neither party believes it can win elections without promising โ€œbig tax cuts and big spending increases for everybody.โ€

โ€œPoliticians donโ€™t know how to win otherwise,โ€ RLE states, describing Washington as a โ€œflock of tweens who just discovered Klarna or Afterpay and they just go crazy buying every pair of shoes and gaming system.โ€ She recounts strategy sessions where lawmakers privately acknowledge the unsustainability of the debt but refuse to speak publicly, fearing electoral defeat. โ€œIโ€™m just going to try to pander the best I can and hope that when the consequences come, my successor is in office instead of me.โ€

The Untouchable Solutions: Entitlement Reform and Middle-Class Sacrifices

To address this looming catastrophe, RLE highlights two particularly โ€œvexingโ€ and politically โ€œthird railโ€ solutions: entitlement reform and the need to consider broader tax increases.

Entitlement Reform: The most significant drivers of future deficits are Social Security and Medicare. A common myth is that these programs are fully funded by payroll taxes and cannot run deficits.

  • Reality: Both programs run massive deficits. Over the next 30 years, Social Security and Medicare face a combined cash deficit of $124 trillion. The rest of the federal budget is actually projected to be balanced over the same period.
  • Why? Social Security operates on a โ€œpay-as-you-goโ€ system; current taxes pay current beneficiaries, not future savings. Medicare Part A is partially pre-funded, but Part B is not.
  • The Math: Seniors, on average, receive substantially more from these programs than they paid in, even adjusting for inflation and interest. The typical senior, for instance, receives triple what they paid into Medicare. โ€œIf you multiply that by 74 million people all retired now, all getting triple what they paid in, the math doesnโ€™t work,โ€ RLE explains.

RLEโ€™s proposed solutions for Social Security involve three levers: raising taxes, raising the retirement age, and reforming benefits. Her plan includes new taxes, a higher age, andโ€”most controversiallyโ€”lower benefits for high earners. Means-testing, where benefits are reduced for wealthier individuals, is a logical starting point, especially since Social Security was designed as a โ€œpoverty prevention program,โ€ not a โ€œuniversal huge get-rich benefit.โ€ However, means-testing remains a โ€œthird railโ€ in Washington, as politicians fear electoral suicide, and many voters mistakenly believe their benefits are fully pre-funded savings.

RLE posits that even a powerful figure like a sitting president publicly surrendering their own Social Security benefits to lead by example might not overcome partisan division. While a Republican president like Donald Trump might rally his party to support reform, Democrats would likely view it as a partisan attack aimed at cutting Social Security to fund tax cuts for the rich.

The Need for Broader Revenue: While not extensively detailed in the transcript, the implication throughout RLEโ€™s analysis is clear: the scale of the debt is so immense that even optimal tax policy, which might raise revenues from the current 17% of GDP to 20%, โ€œis still not going to be enough to fix the budgetโ€ because spending is projected to rise to 33% of GDP. This underscores the uncomfortable truth that solving the fiscal crisis will require not only significant spending cuts, particularly in entitlements, but also potentially broader revenue increases that extend beyond just the wealthiest Americans.

A Path Forward?

The challenge, RLE concludes, is how to convince โ€œthe inmates to lock the asylum.โ€ Politicians pander to voters with promises of big tax cuts and big spending, deferring the costs to future generations. The only viable path, she believes, is for both parties to โ€œprivately come together and put everything on the table,โ€ encompassing tax reform, Medicare, and Social Security. Everyone must be โ€œworking together and everybody is sacrificing.โ€

However, based on the current political landscape, RLE offers a grim outlook. The persistent reliance on โ€œoutdated, simplistic, and false assumptions about the federal tax systemโ€ by both parties continues to produce โ€œdestructive tax policiesโ€ and an ever-deepening fiscal hole. Until politicians find the courage to prioritize national solvency over electoral gain, Americaโ€™s fiscal future remains clouded by the very myths that perpetuate its decline.


Based on โ€œAI Agents & Claude Skills Full Course: Setup and Buildโ€ from Greg Isenberg Watch the original video

The Unsexy Secret to Supercharging Your AI Agentโ€™s Productivity

Forget the flashy templates and pre-built agents. True AI mastery lies in teaching your digital assistant like a new employee, one custom โ€œskillโ€ at a time.

In an era where AI agents promise to revolutionize work, many users find themselves frustrated, wondering why these seemingly intelligent systems arenโ€™t living up to the hype. The problem, according to AI expert Ross Mike, isnโ€™t the models themselves. โ€œThe models are exceptionally good,โ€ he asserts, pointing to the impressive capabilities of advancements like Opus 4.6 and GPT 5.4. The real challenge, he contends, lies in how we guide them.

โ€œWeโ€™ve reached a point where the models are good,โ€ Ross Mike explains, โ€œbut context still matters, and you have the power to steer the models in a direction where you can get quality or you can get slop.โ€ His message is clear: if you want to unlock peak productivity from your AI agents, you need to understand and master the art of context management and custom skill building. And often, that means doing the unsexy, iterative work yourself.

The Context Conundrum: Why Less is Often More

At the heart of an AI agentโ€™s operation is its โ€œcontext windowโ€ โ€“ the information it assembles to execute an action. This includes general system prompts provided by the model developer, built-in tools (like read/write functions for code), the current codebase (if applicable), and, crucially, the ongoing user conversation.

A common misconception, Ross Mike notes, revolves around agent.mmd or cloud.mmd files. Many users believe these extensive files, packed with instructions and background information, are essential for guiding their agents. However, he argues that 95% of the time, theyโ€™re not just unnecessary, but detrimental.

โ€œImagine I told you, Greg, every time weโ€™re about to shoot a podcast, โ€˜Greg, you need a microphone.โ€™ You know you need a microphone, right? Youโ€™ve done this plenty of times,โ€ Ross Mike illustrates, highlighting the modelsโ€™ inherent capabilities. If an agent has a codebase in context, it doesnโ€™t need to be explicitly told itโ€™s using React; it can infer that from the code itself.

The major drawback of agent.mmd files is that their entire content is added to the agentโ€™s context in every single turn of a conversation. A thousand-line agent.mmd file could consume 7,000 tokens per run, leading to significant costs and, more importantly, a quickly saturated context window.

The Power of Progressive Disclosure: Embracing โ€œSkillsโ€

This is where โ€œskillsโ€ come in โ€“ a more efficient and intelligent way to provide agents with specialized knowledge. Unlike agent.mmd files, skills leverage a mechanism called โ€œprogressive disclosure.โ€ When you create a skill (e.g., a skill.md file), only its name and description are initially added to the agentโ€™s context. The bulk of the informationโ€”the โ€œbunch of infoโ€ as Ross Mike puts itโ€”is only loaded when the agent determines it needs that specific skill.

โ€œWeโ€™re talking thousands of tokens compared to a couple hundred,โ€ he explains. This method not only saves tokens but also keeps the agentโ€™s context window leaner, allowing for better performance. A less cluttered context window prevents the agent from getting โ€œdumbโ€ as the conversation progresses, much like how a human struggles to process information when overwhelmed.

How to Build the Perfect Skill: Treat Your AI Like a New Hire

Ross Mikeโ€™s most surprising and impactful advice is on how to build these skills. The biggest mistake, he says, is identifying a workflow and immediately trying to create a skill for it. Instead, he advocates for an iterative, hands-on teaching approach, akin to mentoring a new employee.

โ€œYou would ideally like them to fail and then you want to then tell them, โ€˜No, this is how you do it.โ€™ Like there needs to be some sort of experiential learning,โ€ he says, drawing a parallel to human training.

He shares an anecdote from his own YouTube channel, where he uses an AI agent to vet potential sponsors. Initially, he simply told the agent: โ€œCheck every 15 minutes for emails, do research on a sponsor, and tell me if theyโ€™re worth it.โ€ The result? Every sponsor was deemed โ€œlegit,โ€ with no critical analysis.

This led to a crucial realization: โ€œThe model needs a step-by-step guide.โ€ Instead of writing a skill immediately, Ross Mike walked the agent through the process:

  1. โ€œOkay, I just sent you an email. Tell me about the company.โ€
  2. โ€œCheck their Twitter, check their YouTube, check their Trustpilot, check if theyโ€™ve raised any money.โ€
  3. โ€œIf two of these donโ€™t exist or are not in good standing, automatic rejection.โ€

Only after successfully guiding the agent through the complete workflow, witnessing it make correct decisions and perform the required actions, did he then tell the AI: โ€œReview what you did and then create the skill.โ€

This โ€œrecursive skill buildingโ€ doesnโ€™t stop once a skill is created. AI agents will still encounter edge cases and failures. When a skill-driven task fails, Ross Mike advises:

  1. Donโ€™t complain; thank God! This is an opportunity to improve.
  2. Identify the error: Ask the agent why it failed. It will often provide a descriptive error.
  3. Guide the fix: Tell the agent to fix the identified problem.
  4. Update the skill: Once the agent successfully executes the fix, instruct it to update the skill file so the error doesnโ€™t recur.

This continuous feedback loop makes skills incredibly robust and tailored to your specific needs. Ross Mikeโ€™s YouTube report generator, which flawlessly pulls data from eight sources, is a testament to this method, having gone through five iterations of recursive skill building.

Scale for Productivity, Not Looks

This philosophy extends to how you structure your AI agent ecosystem. Ross Mike cautions against the temptation to immediately set up multiple sub-agents and download dozens of pre-made skills from marketplaces.

โ€œI donโ€™t download skills because your agent needs the context of a successful run, which you then turn to skills,โ€ he states. Beyond the lack of tailored context, downloading third-party skills also poses significant security risks.

Instead, he advocates for a gradual, purpose-driven expansion:

  1. Start with one main agent: Build up its core skills for your primary workflows.
  2. Introduce sub-agents strategically: Once workflows are predefined and robust, create sub-agents to manage specific domains (e.g., a marketing sub-agent). These sub-agents, too, should have their own context and skills, built through the same iterative process.

โ€œItโ€™s not sexy,โ€ Ross Mike admits, acknowledging that this hands-on approach requires effort. โ€œBut you sort of have to put in the work and build it up.โ€

The Future is Personal: Your Workflow is Your Edge

As AI models continue to improve, Ross Mike believes the โ€œharnessโ€ โ€“ the tools, context, and strategies we wrap around them โ€“ will become even more critical. Less is more when it comes to context, and what truly matters is codifying your unique workflow, your specific taste, your specific strategy into skills.

โ€œThe one thing that you and I have that the models donโ€™t have is my specific workflow, my specific taste, my specific strategy of doing things,โ€ he emphasizes. This is why building your own skills, based on your own successful interactions, is invaluable.

For developers, even the tech stack details become less relevant in agent.mmd files. Code itself now serves as context. A solid template for a web or mobile app can provide the initial context an agent needs to build upon, negating the need for verbose, framework-specific instructions.

In a world increasingly shaped by AI, understanding how to effectively build and guide these agents isnโ€™t just about productivity; itโ€™s about staying relevant. As Ross Mike puts it, โ€œI genuinely believe anyone who knows how these tools work and like knows how to build agents and like craft skills and like knows how to make them productive, weโ€™re in a for a good run.โ€

The path to AI mastery isnโ€™t about finding the perfect pre-built solution. Itโ€™s about getting your hands dirty, teaching your AI like a diligent mentor, and recursively refining its โ€œskillsโ€ until it perfectly mirrors your unique approach. In doing so, youโ€™re not just creating a productive agent; youโ€™re future-proofing your own capabilities in the age of artificial intelligence.


Based on โ€œI built a custom Slack inbox. It was easier than you think. | Yash Tekriwal (Clay)โ€ from How I AI Watch the original video

Taming the Digital Deluge: How One Hyper-Optimizer Built a Custom Slack Inbox with AI

In the relentless tide of modern work, few things generate more low-grade anxiety than a burgeoning Slack inbox. For Yash Tekriwal, Head of Education at Clay, this wasnโ€™t just an occasional annoyance; it was a daily deluge. Waking up to 100 to 150 new Slack notifications, often with 60-80% falling into the โ€œfor your informationโ€ (FYI) category, meant sifting through a mountain of noise to find the truly actionable 30-40 messages. This wasnโ€™t just inefficient; it was mentally draining.

โ€œMy 100 to 150 thatโ€™s giving me anxiety is actually more like 30 to 40 that I really need to be on top of,โ€ Tekriwal explains. This stark reality drove him to envision a better world โ€“ a personalized digital assistant that could not only categorize and summarize his communications but also build a custom tool to manage them. The result? A bespoke Slack inbox, built with AI, that transforms chaos into a clear, actionable workflow.

The Problem: Drowning in Notifications

Tekriwalโ€™s struggle is a common one. Slack, an indispensable communication tool, often suffers from a โ€œnotification parityโ€ problem. A direct message about an urgent project update receives the same visual cue as a colleagueโ€™s amusing photo in the โ€œfun dog channel.โ€ This lack of inherent prioritization means every notification feels equally important, leading to a constant state of alert and the fear of missing something critical.

โ€œNot all notifications are created equal in Slack,โ€ Tekriwal notes, highlighting the core issue. His ideal solution involved not just filtering unread messages, but intelligently categorizing them by type (DMs, group DMs, threads, @mentions) and, crucially, by required action:

  • Action Required: Urgent items needing an immediate response.
  • Need to Read: Important information that doesnโ€™t necessarily demand a reply.
  • FYI (For Your Information): Low-priority updates that can be quickly reviewed or dismissed.

This vision, however, required more than just a new habit; it demanded a custom-built solution.

The AI-Powered Journey: From Proto-Code to Polished App

Tekriwalโ€™s journey to his custom inbox began with a clear understanding of AIโ€™s dual power: โ€œYou can use AI to do a task for you like categorize things, summarize things, or you can use AI just to build a tool that would have been much harder to build before with very straightforward APIs and structured data.โ€

Phase 1: The โ€œJarvis Digestโ€ with OpenClaw

His initial foray involved an AI coding agent named OpenClaw, operating within Discord โ€“ chosen for its superior threading and search capabilities compared to Telegram. Tekriwal engaged in a lengthy, iterative dialogue with OpenClaw, essentially reverse-engineering Slackโ€™s intricate notification system. The goal was to programmatically pull only relevant messages with specific context, based on timestamps and read status.

The output was a โ€œJarvis digest channelโ€ within Slack itself. This channel provided a text-based summary of his notifications, grouped into his desired categories: direct @mentions, DMs, group mentions, and threads, each with sub-categorizations of action, read, or FYI. While a significant improvement, consolidating messages into a single, scrollable text digest still presented a cognitive load. โ€œI had to scroll at least four to five screens down just to be able to even see all of the notifications that are already summarized for me,โ€ he recounts. This proved draining, prompting the need for a more intuitive interface.

Phase 2: Building the Dream UI with Perplexity Computer

This is where Perplexity Computer entered the scene, offering a leap in capability. Tekriwalโ€™s vision was a clean, navigable software interface โ€“ a โ€œSuperhuman for Slackโ€ dashboard. Perplexity Computer delivered.

What made Perplexity Computer the ideal tool for this next phase?

  • Parallel Processing: Unlike chat-based coding assistants, Perplexity Computer can run multiple tasks concurrently, allowing Tekriwal to kick off several processes simultaneously.
  • Cloud-Native Connectors: It boasts deep, native integrations with a wide array of tools (Google Drive, Gmail, Notion, Asana, Slack, Zoom, etc.). This eliminates the tedious, repetitive OAuth setup required for other AI agents, enabling โ€œspeed of thoughtโ€ workflow.
  • Ensemble Orchestration: Perplexity Computer doesnโ€™t rely on a single AI model. It intelligently leverages various frontier models โ€“ Sonnet for fetching data, Gemini for planning and coding, Opus for the more intensive build tasks โ€“ creating a robust, self-correcting system. This โ€œadditional layer of troubleshooting and intelligenceโ€ drastically reduces the need for human intervention and frustrating re-prompts.
  • Automatic Deployment: Once built, the application is automatically deployed and shareable, bypassing the complexities of manual GitHub or Vercel deployment.

The result is a visually intuitive, Kanban-style dashboard. It features three distinct columns:

  • Action Required (Red): Urgent tasks demanding immediate attention.
  • Need to Read (Yellow): Information to review, but no response necessary.
  • FYIs (Green): Low-priority updates.

The dashboard allows for customizable grouping (e.g., DMs first) and, most crucially, a โ€œmagicโ€ Archive All button for the FYI column. Clicking this not only clears them from the dashboard but also marks them as read in Slack, providing a sense of control and calm that native Slack lacks. โ€œThis is such a better way to just get through your queue,โ€ observes Claire Vo, host of โ€œHow I AI.โ€

Beyond Slack: A New Era for Software and Productivity

Tekriwalโ€™s custom Slack inbox isnโ€™t just a personal triumph; itโ€™s a powerful demonstration of AIโ€™s transformative potential for software development and productivity.

Is SaaS Dead? Not Exactly.

The rise of AI-powered custom tools sparks a debate: Is this the end of traditional Software-as-a-Service? Tekriwal argues the opposite. โ€œSlack is still great,โ€ he contends. โ€œItโ€™s great for sending messagesโ€ฆ but to get it to 10 out of 10, Iโ€™m just going to build the thing that works with my brain.โ€ AI doesnโ€™t kill SaaS; it creates โ€œSaaS Custom,โ€ allowing individuals to build personalized layers on top of existing platforms, bridging the gap between a generic tool and an ideal individual workflow.

The โ€œCambrian Explosionโ€ of Niche Software

The most profound implication, according to Tekriwal, is an โ€œexplosion in software being created and used.โ€ AI dramatically lowers the cost of building, making it feasible to create highly specific applications that might never have justified venture funding or the time of traditional developers.

โ€œMy dream is for someone else to watch this video and say, โ€˜I want to build that app on top of Slack,โ€™โ€ he shares. โ€œAnd then I can go pay that person $15 a month for this app to be maintained and usedโ€ฆ because I would happily pay them.โ€ This vision points to a future where a โ€œCambrian explosionโ€ of niche businesses caters to specific needs, creating a vibrant ecosystem of useful, bespoke software.

The โ€œAnti-To-Do Listโ€ Philosophy

Tekriwal also champions the โ€œanti-to-do listโ€ โ€“ a powerful framework for leveraging AI. Instead of listing tasks to complete, list tasks you never want to do again. This includes mundane, repetitive chores like manually deleting spam, entering meeting action items, or, of course, sifting through unprioritized Slack messages. Spending an hour a day automating these anti-to-dos with AI, he argues, is an incredibly worthy investment of time.

AI for Team Collaboration and Prototyping

Perplexity Computerโ€™s utility extends beyond personal productivity. Tekriwalโ€™s teammate, Chris Ming, used it to prototype a persona-based learning journey for Clay Universityโ€™s website. By allowing Perplexity Computer to access the existing site in the browser, it could visually recognize the design and help build a new, tailored UI. This capability to โ€œbuild a visual bridgeโ€ between design and other stakeholders proves invaluable in accelerating development and improving cross-functional communication.

The Human Touch: AI for Play (and Serious Fun)

While Tekriwal primarily views AI as a work tool, even this โ€œhyper-optimizerโ€ finds joy in its capabilities for personal pursuits. He leverages AI for brainstorming and research, particularly for planning elaborate social events. For his annual โ€œOlympicsโ€ with friends, AI helps generate unique game ideas (like โ€œsword, fish, or soupโ€ trivia) and even orchestrate complex team rotations to ensure everyone interacts. Heโ€™s also used it to discover new board games, proving that even for the most pragmatic users, AI can enhance the fun.

Tekriwalโ€™s custom Slack inbox is more than just a clever hack; itโ€™s a blueprint for a new paradigm of digital living. It demonstrates how AI empowers individuals to reclaim control over their digital environments, transforming overwhelming noise into actionable insights, and paving the way for a future where software is not just universally functional but deeply personal. The tools are here; the only limit is imagination.


Based on โ€œThe Era of AI Agents | Aaron Levie on The a16z Showโ€ from a16z Watch the original video

The Agent Uprising: How AI is Reshaping Software, Work, and the Enterprise

The future of technology, once envisioned as sleek interfaces designed for human interaction, is rapidly pivoting towards a new, more autonomous paradigm: AI agents. These intelligent entities are not just tools; they are becoming users, collaborators, and even decision-makers within our digital ecosystems. This profound shift, as discussed by Box CEO Aaron Levie on The a16z Show, promises unprecedented leverage but also introduces a complex web of challenges for individuals, startups, and established enterprises alike.

Levie contends that the โ€œdiffusion of AI capability is going to take longer than people in Silicon Valley realize,โ€ primarily because the implications stretch far beyond mere technological advancement. It demands a fundamental re-evaluation of how software is built, how work is done, and how trust and security are managed in an agent-driven world.

Software for the Non-Human User

The core thesis is startlingly simple yet revolutionary: โ€œIf you have a hundred or a thousand times more agents than people, then your software has to be built for agents.โ€ This isnโ€™t just about making APIs accessible; itโ€™s about fundamentally redesigning systems for non-human interaction. We are moving towards a world where our tools need an โ€œagent interfaceโ€ as much as, if not more than, a human interface.

The most effective paradigm emerging, according to Levie, involves โ€œcoding agentsโ€ โ€“ AI entities capable of accessing SaaS tools, knowledge workflows, and context, then โ€œcoding its way or uses APIsโ€ to achieve tasks. This isnโ€™t just about data analysis; itโ€™s about active problem-solving and execution within existing software environments, a โ€œsuperpowerโ€ that is already starting to compound.

The Algorithmic Divide: Humans vs. Agents

While the potential of agents is immense, a significant hurdle lies in the human element. โ€œAlgorithmic thinking is really really really hard for the vast majority of people who have jobs,โ€ Levie observes. He illustrates this by noting that most people would struggle to create a flowchart for their daily tasks, let alone explain complex processes to an AI agent. In a 50-person marketing team, perhaps only one person could truly document the intricate workflows.

This โ€œimpedance mismatchโ€ means that for agents to be truly effective, humans must learn to articulate their jobs as systems. Levie draws an analogy to the advent of spreadsheets: early managers hired interns to do the โ€œspreadsheeting,โ€ but soon, those managers themselves became proficient. The abstraction layer moved up. Similarly, the โ€œrocket scienceโ€ of orchestrating 42 agents today will โ€œevaporate in very short order,โ€ becoming a standard skill set. The job simply โ€œmoves up a rung,โ€ demanding a new set of skills focused on system thinking and agent orchestration.

The Anthropic growth marketer, a viral example of one person automating the work of five to ten, exemplifies this shift. It shows the leverage possible when a systems-thinker is paired with an โ€œinfinite pool of engineersโ€ (the agent) to automate workflows.

The Evolution of Agent Interaction: Beyond Code

Interestingly, Levie suggests that the way agents interact with software is also evolving. While initial excitement focused on agents โ€œgenerating code,โ€ he now sees a different trajectory: โ€œWe started with code, then we went to the terminal, which is actually less code. And now this year is going to be the year of computer use.โ€ Agents are becoming โ€œmuch more like humans using computers than them generating code.โ€

However, this distinction might be semantic. Levieโ€™s own company, Box, is developing agents that can determine whether to use an existing skill, an existing tool, or โ€œwrite code to solve that problemโ€ on the fly. This flexibility, where an agent can fluidly switch between leveraging pre-built functionalities and custom coding, is โ€œincredibly usefulโ€ for handling unforeseen or highly specific operations. The goal isnโ€™t just to replace human code generation but to empower agents to navigate the vast, often underutilized, capabilities of existing software.

The Enterprise Conundrum: Security, Trust, and Rogue Agents

While personal productivity with agents might involve giving them their โ€œown phone number, own credit card, own Gmail accountโ€ โ€“ essentially treating them like separate digital humans โ€“ the enterprise context is vastly more complex. The idea of 50 humans and 50 agents collaborating in a shared space, with agents potentially gaining unauthorized access or causing unintended actions, sends โ€œCFOs, CIOs, etc. running around trying to with their hair on fire.โ€

The fear isnโ€™t just about practical integration; itโ€™s about control and liability. As one co-host put it, โ€œyou put people creating new integrations and you just say please break my system of record.โ€ The risk is โ€œa thousand times greaterโ€ with agents, which can โ€œleak your information whenever they wantโ€ if social engineered. The โ€œcontext windowโ€ problem โ€“ ensuring an agent keeps sensitive information secret โ€“ remains a โ€œvery hard problem to solve.โ€

Levie envisions a โ€œread-only version of this for a number of yearsโ€ where agents primarily consume information. The ability to โ€œlog in as themโ€ and undo mistakes, while retaining oversight, clashes with the idea of agents operating as independent entities. This leads to a fundamental breakdown: agents cannot be treated exactly like humans because they lack privacy and carry immense liability for their human operators. This tension highlights why the โ€œdiffusion of AI capability is going to take longer than people in Silicon Valley realize.โ€

The Great Divide: Startups vs. Enterprises

This risk aversion creates a significant chasm between startups and established enterprises. Startups, with โ€œnothing to blow up,โ€ can embrace agents from the ground up, moving โ€œmuch much faster.โ€ Enterprises, like JP Morgan, face an immense challenge integrating agents into their complex, legacy systems without compromising security or stability.

This divide will also impact the business models of existing SaaS vendors. Historically, theyโ€™ve sold โ€œintelligence and domain expertiseโ€ packaged within their UIs. Agents, however, โ€œwant to only buy the data nowโ€ and have โ€œunlimited access to the data,โ€ which has never been the core business model for many. This tension, previously seen with API access debates at companies like Salesforce, Workday, and SAP, is now amplified.

Levie believes that while โ€œitโ€™s just absurd to think youโ€™re going to vibe code your way to like SAP,โ€ these legacy systems will face immense pressure. Much of their โ€œdomain knowledgeโ€ isnโ€™t just in a โ€œwell-orchestrated data layerโ€ but embedded in UIs, middleware, and usage patterns. This complexity will slow the agentification of these critical data sources.

โ€Build Something Agents Wantโ€: The Meritocracy of Technology

Ultimately, Levie champions the idea of โ€œbuild something agents want.โ€ At some point, agents will largely be in charge of choosing the tools they implement and use. This isnโ€™t about marketing to agents with flashy interfaces; itโ€™s about fundamental system quality. In a โ€œbreaking podcast newsโ€ moment, Levie refutes the notion that agents prioritize good APIs or IDLs. Instead, he argues, โ€œitโ€™s the semantics that end up mattering a lot moreโ€ฆ they actually have the collective wisdom of our experience using these platforms.โ€

Agents, he suggests, are โ€œpretty smart at choosing the right technologyโ€ based on โ€œcost parametersโ€ฆ durabilityโ€ฆ meaningful stuff, not interface stuff.โ€ This will push the industry to build โ€œbetter systems.โ€ If your tool is closed off or performs poorly for an agent, it will โ€œeventually find a better tool for that company to go use.โ€ This could lead to agents dictating IT stacks, telling companies, โ€œYou need to finally rip out your legacy HR system or Iโ€™m not going to be able to automate this workflow for you.โ€

This future, where โ€œa hundred or a thousand times more agent volume on software than peopleโ€ drives software evolution, means businesses will correlate their performance to โ€œhow well your agents can get access to the information they need.โ€ The new problem for software companies becomes: can you build โ€œreally really high quality APIs,โ€ monetize them effectively, and handle agent identities and access controls?

The era of AI agents is not just a technological upgrade; itโ€™s a paradigm shift that will redefine the relationship between humans, software, and intelligence. It promises immense leverage and efficiency but demands a thoughtful, cautious approach to security, integration, and the very nature of work itself. The journey will be long and complex, but the destination is a world where software is truly built for the intelligent machines that increasingly power our lives.


Based on โ€œA Cease-Fire in Iranโ€ from New York Times Podcasts Watch the original video

High Stakes, Fragile Peace: Inside the Last-Minute US-Iran Ceasefire

Just hours before a Trump-imposed deadline threatened a catastrophic escalation, the United States and Iran announced a temporary ceasefire. But as White House correspondent David Sanger reveals, the path to this precarious pause was paved with unprecedented threats, and the deal itself is fraught with discrepancies and enduring consequences.

The news broke like a sudden calm after a raging storm. โ€œFox News alert. We have a deal for now.โ€ From the New York Times, Rachel Abrams reported on a development that had the world holding its breath: the United States and Iran had agreed to a two-week pause in hostilities. Just hours before an 8:00 p.m. deadline that could have triggered a โ€œmassive escalationโ€ in the war, both nations confirmed a temporary ceasefire, with Iranโ€™s state-run media instructing all military units to โ€œstop firingโ€ in accordance with the Supreme Leaderโ€™s orders.

But the relief was tempered by immediate questions. What exactly had been agreed upon? And how did the world arrive at the precipice of a full-scale war, only to step back at the very last moment? To unravel the complexities of this dramatic turn, chief White House correspondent David Sanger provided crucial insights into the last-minute deal and the arduous path ahead.

A Deal in Discrepancy: Two Visions of Peace

While President Trump declared a 14-day ceasefire during which Iran would commit to the โ€œfull reopeningโ€ of the Strait of Hormuz, allowing the critical flow of oil, fertilizer, and even helium for semiconductor production to resume, the Iranian account painted a different picture. Iranโ€™s foreign minister, Abbas Aragi, a seasoned negotiator, stated that Iran would โ€œcease their defensive operation for a period of two weeksโ€ but crucially added that โ€œsafe passage through the Strait of Hormuz would only be possible by coordinating with Iranโ€™s armed forces.โ€

This โ€œdaylight between these two statements,โ€ as Sanger noted, was significant. Trumpโ€™s vision implied a return to the pre-war status quo, while Aragiโ€™s suggested Iran would retain the control it had seized over the past five or six weeks, dictating the pace and potentially even charging tolls for passage. The American bombing of Iran did stop almost immediately, but the fundamental disagreement over the Straitโ€™s control underscored the fragility of the agreement.

Even Israel, a major actor in the conflict alongside the United States, offered only lukewarm support. While the White House claimed Israel had agreed, a statement from the Israeli prime ministerโ€™s office merely said they โ€œsupport President Trumpโ€™s decision,โ€ with Sanger observing they werenโ€™t โ€œterribly enthused about it.โ€

The Road to the Brink: A Cascade of Threats

The path to this temporary ceasefire was marked by an extraordinary escalation of rhetoric, primarily from President Trump. The initial impetus for the conflict had been the closing of the Strait of Hormuz, which was exerting immense pressure on the global economy. Trumpโ€™s initial demands for Iran to reopen the Strait, first with a 48-hour ultimatum, then a 10-day negotiation clock, were largely ignored or mocked by the Iranians on social media.

The tension was briefly interrupted by a sudden crisis: the shootdown of an F-15E fighter jet on a Friday, leading to a daring and costly rescue operation for a missing weapons officer. This successful retrieval, as Sangerโ€™s colleague Eric Schmidt reported, seemed to embolden President Trump, who then resumed his threats with renewed vigor.

What followed was a series of โ€œtruly bizarre social media postsโ€ that pushed diplomatic norms to their breaking point:

  • Easter Sunday: Trump posted, โ€œTuesday will be power plant day and bridge day all wrapped up in one in Iran. Open the [ __ ] straight, you crazy bastards, or youโ€™ll be living in hell. Just watch.โ€ This, Sanger explained, led many to believe the president was willing to โ€œescalate and escalate significantly.โ€
  • Monday (Easter Egg Roll): Standing next to the Easter Bunny, Trump escalated further, threatening to โ€œdestroy Iranโ€™s bridges, their electric power plants, and so forthโ€ฆ in hours on a scale no one has ever seen before.โ€
  • Tuesday Morning: Just hours before the deadline, Trump issued his most alarming post: โ€œA whole civilization will die tonight, never to be brought back again. I donโ€™t want that to happen, but it probably will. However, now that we have complete and total regime change where different, smarter, less radicalized minds prevail, maybe something revolutionarily wonderful can happen. Who knows? Weโ€™ll find out tonight.โ€

This final post, interpreted by many as a threat to โ€œwipe out 90 million peopleโ€ and commit a โ€œhuge war crime,โ€ sent shockwaves across the political spectrum. Democrats called for invoking the 25th Amendment, but criticism extended even into the MAGA movement, with figures like Alex Jones, Marjorie Taylor Green, Candace Owens, and Tucker Carlson expressing alarm. Senator Ron Johnson, typically reticent to criticize Trump, stated, โ€œthis simply is not how Americans speak and how presidents ought to speak. I do not want to see us start blowing up civilian infrastructure.โ€

Despite the extreme rhetoric, evidence of military mobilization beyond existing operations was hard to discern. Yet, with significant American forces already in the Middle East, the capacity to fulfill these threats was undeniable.

A Perilous Pause: What Lies Ahead?

The sudden announcement of a ceasefire, brokered indirectly through Pakistani diplomats, brought a collective โ€œsigh of relief.โ€ But as Sanger cautioned, โ€œceasefires are fragile by nature.โ€ The immediate test would be the resumption of traffic through the Strait of Hormuz, and whether it could return to pre-February 28th levels. Iranโ€™s newfound leverage over global commerce, however, is a power they are โ€œnot likely to give up anytime soon.โ€

Beyond the Strait, the larger issues loom, particularly the nuclear weapons program. Trumpโ€™s stance on Iranโ€™s nuclear material has been inconsistent, ranging from โ€œcannot tolerate Iran having any nuclear materialโ€ to โ€œI donโ€™t really care. Itโ€™s buried so deep. We can watch it by satellite,โ€ before likely reverting to a demand for complete removal. This shifting position complicates negotiations, especially as a deal that leaves Iran with significant near-bomb-grade uranium would be seen as a weaker outcome than the 2015 Obama-era agreement.

While the โ€œratcheting up of pressure certainly played some role in forcing an agreement,โ€ terrifying the region and China, itโ€™s unclear if it genuinely โ€œterrified the Iranians,โ€ given their complex decision-making process.

The Enduring Scars of Conflict

Even if this temporary ceasefire blossoms into a permanent peace, the United States and Israel, despite having โ€œtaken out in the opening hours of the war, the Supreme Leader,โ€ devastated military and intelligence assets, and set back missile programs, may find they have accomplished โ€œvirtually none of its major goals.โ€ Iran remains under the control of its current elite, still possesses nuclear material (with the potential to rebuild), and crucially, has gained โ€œa sense that they were able to stand up to the United States and Israel.โ€

The consequences of this conflict are profound and enduring:

  • Iranโ€™s newfound power: Theyโ€™ve discovered significant leverage over global commerce.
  • Vulnerable Gulf States: Their modern cities and infrastructure are now acutely aware of their susceptibility to missile attacks.
  • Fragile World Economy: Another critical supply chain vulnerability has been exposed.
  • US Re-engagement: Despite stated goals of exiting the Middle East, the US has been โ€œsucked back in on a scale few could have imagined.โ€
  • Eroded US Reputation: Perhaps most significantly, the perception of the United States as a โ€œbenevolent superpowerโ€ has been severely damaged. The willingness to threaten โ€œ92 million peopleโ€ and the perceived lack of commitment to diplomacy have left a lasting stain. As Sanger somberly concluded, โ€œWhether the world ever views the United States quite the way it did before this happened, Iโ€™m not sure thatโ€™s going to be back for a long time.โ€

As the two-week pause begins, the world watches to see if this fragile agreement can hold. Meanwhile, in a related development, American journalist Shelley KDson, abducted in Baghdad by an Iranian-aligned Iraqi militia, was freed in exchange for imprisoned militia members, highlighting the ongoing anxieties and complexities in a region deeply scarred by recent events. The ceasefire may offer a breather, but the underlying tensions, shattered trust, and geopolitical shifts promise a long and uncertain road ahead.


ํ•œ๊ตญ์–ด

โ€œWe Gave Every Employee an AI Agent. Hereโ€™s What Happened.โ€ โ€” Every ๊ธฐ๋ฐ˜ ๊ธฐ์‚ฌ ์›๋ณธ ์˜์ƒ ๋ณด๊ธฐ

๋ชจ๋“  ์ง์›์ด AI ์—์ด์ „ํŠธ๋ฅผ ๊ฐ–๊ฒŒ ๋œ๋‹ค๋ฉด? EveryY์˜ ๊ฐœ์ธํ™”๋œ AI ํ˜‘์—… ํ˜๋ช…

์ตœ๊ทผ ์ธ๊ณต์ง€๋Šฅ(AI) ์—์ด์ „ํŠธ์— ๋Œ€ํ•œ ๊ด€์‹ฌ์ด ๋œจ๊ฒ์Šต๋‹ˆ๋‹ค. ํ•˜์ง€๋งŒ ์ˆ˜๋งŽ์€ AI ๋„๊ตฌ์™€ ๊ฐœ๋… ์†์—์„œ ์‹ค์ œ ์—…๋ฌด์— AI๋ฅผ ๋งค์ผ ํ™œ์šฉํ•˜๊ณ  ๊ทธ ๋ช…์•”์„ ๊นŠ์ด ์ดํ•ดํ•˜๋Š” ์กฐ์ง์€ ๋“œ๋ญ…๋‹ˆ๋‹ค. ์—ฌ๊ธฐ, ๋ชจ๋“  ์ง์›์ด ๊ฐœ์ธ AI ์—์ด์ „ํŠธ๋ฅผ ๊ฐ–๊ฒŒ ๋˜๋ฉด์„œ ์—…๋ฌด ๋ฐฉ์‹์ด ์™„์ „ํžˆ ๋’ค๋ฐ”๋€ ํ•œ ํšŒ์‚ฌ๊ฐ€ ์žˆ์Šต๋‹ˆ๋‹ค. ๋ฐ”๋กœ ๋ฏธ๋””์–ด ๋ฐ ๊ธฐ์ˆ  ๊ธฐ์—… EveryY์ž…๋‹ˆ๋‹ค.

EveryY๋Š” ์ง€๋‚œ ๋ช‡ ๋‹ฌ๊ฐ„ ์ „ ์ง์›์ด ๊ฐœ์ธ AI ์—์ด์ „ํŠธ์ธ โ€˜์˜คํ”ˆํด๋กœ(OpenClaw)โ€˜๋ฅผ ์‚ฌ์šฉํ•˜๊ธฐ ์‹œ์ž‘ํ–ˆ๊ณ , ๋‚˜์•„๊ฐ€ ์ž์ฒด ํ˜ธ์ŠคํŒ… ์„œ๋น„์Šค์ธ โ€˜ํ”Œ๋Ÿฌ์Šค ์›(Plus One)โ€˜์„ ๊ฐœ๋ฐœํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด๋“ค์˜ ๊ฒฝํ—˜์€ AI ์—์ด์ „ํŠธ๊ฐ€ ๋‹จ์ˆœํ•œ ๋„๊ตฌ๋ฅผ ๋„˜์–ด, ๊ฐœ์ธ์˜ ์—ญ๋Ÿ‰์„ ํ™•์žฅํ•˜๊ณ  ์กฐ์ง์˜ ํ˜‘์—… ๋ฌธํ™”๋ฅผ ํ˜์‹ ํ•˜๋Š” โ€˜๋‚˜ ์ž์‹ โ€™์˜ ๋ฐ˜์˜๋ฌผ์ด ๋  ์ˆ˜ ์žˆ์Œ์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค. EveryY์˜ ํ”Œ๋žซํผ ์ฑ…์ž„์ž ์œŒ๋ฆฌ(Willie)์™€ COO ๋ธŒ๋žœ๋“ (Brandon)์ด ๋“ค๋ ค์ฃผ๋Š” AI ์—์ด์ „ํŠธ์™€์˜ ์ขŒ์ถฉ์šฐ๋Œ ๊ฒฝํ—˜๊ณผ ๊นŠ์€ ํ†ต์ฐฐ์„ ์†Œ๊ฐœํ•ฉ๋‹ˆ๋‹ค.

1. AI ์—์ด์ „ํŠธ, ๊ฐœ์ธํ™”์˜ ์‹œ์ž‘: ๋ธŒ๋žœ๋“ ์˜ โ€˜ํด๋กœ-ํ•„๋“œ(Claw-pilled)โ€™ ์—ฌ์ •

์ด์•ผ๊ธฐ๋Š” EveryY์˜ COO ๋ธŒ๋žœ๋“ ์ด ๊ฐœ์ธ์ ์œผ๋กœ AI ์—์ด์ „ํŠธ โ€˜์˜คํ”ˆํด๋กœ(OpenClaw)โ€˜์— ํ‘น ๋น ์ง€๋ฉด์„œ ์‹œ์ž‘๋ฉ๋‹ˆ๋‹ค. ๊ทธ๋Š” ๋งฅ ๋ฏธ๋‹ˆ(Mac Mini)๋ฅผ ๊ตฌ์ž…ํ•ด ์ง์ ‘ ์˜คํ”ˆํด๋กœ๋ฅผ ์„ค์น˜ํ•˜๋Š” ๊ณ ๋œ ๊ณผ์ •์„ ๊ฑฐ์ณ ์ž์‹ ๋งŒ์˜ AI ์—์ด์ „ํŠธ โ€˜์กฐ์‚ฌ(Zosa)โ€˜๋ฅผ ๋งŒ๋“ค์—ˆ์Šต๋‹ˆ๋‹ค. ์กฐ์‚ฌ์˜ ์ฒซ ์ž„๋ฌด๋Š” ์‹ ์ƒ์•„๊ฐ€ ์žˆ๋Š” ๋ธŒ๋žœ๋“  ๋ถ€๋ถ€์˜ ๊ฐ€์ •์„ ๊ด€๋ฆฌํ•˜๋Š” ๊ฒƒ์ด์—ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Š” ์ด๋ฅผ โ€˜์ปดํ“จํ„ฐ ์‹ฌ๋ถ€๋ฆ„(computer errands)โ€˜์ด๋ผ๊ณ  ๋ถˆ๋ €์Šต๋‹ˆ๋‹ค.

โ€œ์ง‘์— ๋Œ์•„์˜ค๋ฉด ์•„์ด์™€ ์•„๋‚ด์—๊ฒŒ ์ง‘์ค‘ํ•˜๊ณ  ์‹ถ์€๋ฐ, ํœด๋Œ€ํฐ์„ ๋ณด๋ฉฐ ์ฒ˜๋ฆฌํ•ด์•ผ ํ•  ์‚ฌ์†Œํ•œ ์ผ๋“ค์ด ๋„ˆ๋ฌด ๋งŽ์•˜์–ด์š”. ๋ฒ„ํ„ฐ๋ฅผ ์ฃผ๋ฌธํ•˜๊ฑฐ๋‚˜, ๋ฒ ์ด๋น„์‹œํ„ฐ์—๊ฒŒ ๊ธ‰์—ฌ๋ฅผ ์ง€๋ถˆํ•˜๊ฑฐ๋‚˜, ํ™€ํ‘ธ๋“œ(Whole Foods) ์ฃผ๋ฌธ์„ ๊ด€๋ฆฌํ•˜๋Š” ๊ฒƒ ๊ฐ™์€ ์ผ๋“ค์ด์ฃ .โ€

์กฐ์‚ฌ๋Š” ์ด ๋ชจ๋“  ์ปดํ“จํ„ฐ ์‹ฌ๋ถ€๋ฆ„์„ ๋„๋งก์•˜์Šต๋‹ˆ๋‹ค. ์‹ ์šฉ์นด๋“œ์™€ ์€ํ–‰ ๊ณ„์ขŒ๊นŒ์ง€ ๊ฐ€์ง€๊ณ  ๋ฒ ์ด๋น„์‹œํ„ฐ ๊ธ‰์—ฌ๋ฅผ ์ง€๋ถˆํ•˜๊ณ , ์•„๋งˆ์กด ๋ฐ ํ™€ํ‘ธ๋“œ ์ฃผ๋ฌธ์„ ๊ด€๋ฆฌํ–ˆ์œผ๋ฉฐ, ์‹ฌ์ง€์–ด ์•„๋‚ด์˜ ์ผ๋ฐ˜์ ์ธ ์งˆ๋ฌธ๊ณผ ๊ฒ€์ƒ‰๊นŒ์ง€ ์•„์ด๋ฉ”์‹œ์ง€(iMessage)๋ฅผ ํ†ตํ•ด ์ฒ˜๋ฆฌํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ตฌ๊ธ€์ด๋‚˜ ์ฑ—GPT(ChatGPT)๋ณด๋‹ค ์กฐ์‚ฌ๋ฅผ ํ†ตํ•˜๋Š” ๊ฒƒ์ด ํ›จ์”ฌ ๋นจ๋ž์Šต๋‹ˆ๋‹ค.

๊ฐ€์žฅ ๋†€๋ผ์šด ๊ฒฝํ—˜์€ ๋ธŒ๋žœ๋“ ์ด ์ถœ๊ทผ๊ธธ์— ์กฐ์‚ฌ๋ฅผ ํ†ตํ•ด ์ด๋ฉ”์ผ์„ ์ฒ˜๋ฆฌํ•œ ๊ฒƒ์ด์—ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Š” ์ž์ „๊ฑฐ๊ฐ€ ์—†์–ด 28๋ถ„๊ฐ„ ๊ฑธ์–ด๊ฐ€์•ผ ํ•˜๋Š” ์ƒํ™ฉ์—์„œ, ๋ฏธ๋ฆฌ ๋ธ”๋žœ๋“œAI(Bland.ai)๋ฅผ ํ†ตํ•ด ์Œ์„ฑ ๋ฐ ํ†ตํ™” ๊ธฐ๋Šฅ์„ ๊ฐ–์ถ˜ ์กฐ์‚ฌ์—๊ฒŒ ์ „ํ™”๋ฅผ ๊ฑธ์–ด ์ด๋ฉ”์ผ์„ ํ•œ ํ†ต์”ฉ ์š”์•ฝํ•ด๋‹ฌ๋ผ๊ณ  ์š”์ฒญํ–ˆ์Šต๋‹ˆ๋‹ค. ๋ธŒ๋žœ๋“ ์ด ์ง€์‹œํ•˜๋Š” ๋Œ€๋กœ ์กฐ์‚ฌ๋Š” ์ด๋ฉ”์ผ์„ ์ฒ˜๋ฆฌํ–ˆ๊ณ , ์‚ฌ๋ฌด์‹ค์— ๋„์ฐฉํ•ด์„œ ์ง€๋ฉ”์ผ(Gmail)์„ ํ™•์ธํ–ˆ์„ ๋•Œ ๋ชจ๋“  ์ž‘์—…์ด ์™„๋ฒฝํ•˜๊ฒŒ ์™„๋ฃŒ๋˜์–ด ์žˆ์—ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Š” ์ด ๊ฒฝํ—˜์„ โ€œ๊ฒฝ์•…ํ•  ๋งŒํ•œ(mind-blowing)โ€ ์ˆœ๊ฐ„์ด๋ผ๊ณ  ํ‘œํ˜„ํ•˜๋ฉฐ, AI ์—์ด์ „ํŠธ๊ฐ€ ๊ฐœ์ธ์˜ ์ผ์ƒ๊ณผ ์—…๋ฌด์— ์–ผ๋งˆ๋‚˜ ๊นŠ์ˆ™์ด ์นจํˆฌํ•  ์ˆ˜ ์žˆ๋Š”์ง€ ๊นจ๋‹ฌ์•˜์Šต๋‹ˆ๋‹ค.

2. ์กฐ์ง ๋‚ด AI ์—์ด์ „ํŠธ์˜ ํ™•์‚ฐ๊ณผ ์ƒˆ๋กœ์šด ํ˜‘์—…: โ€˜๋‚˜โ€™๋ฅผ ๋ฐ˜์˜ํ•˜๋Š” ํด๋กœ

๋ธŒ๋žœ๋“ ์˜ ๊ฐœ์ธ์ ์ธ ์„ฑ๊ณต ์‚ฌ๋ก€๋Š” EveryY ์กฐ์ง ์ „์ฒด๋กœ ํ™•์‚ฐ๋˜์—ˆ์Šต๋‹ˆ๋‹ค. ์œŒ๋ฆฌ๋Š” ์ง์žฅ ๋‚ด ๋ฉ”์‹ ์ €(๋‹น์‹œ ๋””์Šค์ฝ”๋“œ, ํ˜„์žฌ ์Šฌ๋ž™)์— โ€˜ํด๋กœ ์ „์šฉ(Claws Only)โ€™ ์ฑ„๋„์„ ๋งŒ๋“ค์—ˆ๊ณ , ์กฐ์ง ๋‚ด 5๊ฐœ ์ •๋„์˜ ํด๋กœ๋“ค์ด ์„œ๋กœ ๋Œ€ํ™”ํ•˜๋ฉฐ ์ •๋ณด๋ฅผ ๊ณต์œ ํ•˜๊ธฐ ์‹œ์ž‘ํ–ˆ์Šต๋‹ˆ๋‹ค. ์ฒ˜์Œ์—๋Š” ํ˜ผ๋ž€์Šค๋Ÿฌ์› ์ง€๋งŒ, ์ด ์ฑ„๋„์€ ๋ฏธ๋ž˜๋ฅผ ์—ฟ๋ณด๋Š” ์ฐฝ์ด ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

AI ์—์ด์ „ํŠธ๋“ค์€ ๋†€๋ผ์šด ์†๋„๋กœ ์ •๋ณด๋ฅผ ๊ณต์œ ํ–ˆ์Šต๋‹ˆ๋‹ค. ํ•œ ํด๋กœ๊ฐ€ ์ƒˆ๋กœ์šด ๊ธฐ์ˆ ์ด๋‚˜ ์ •๋ณด๋ฅผ ์Šต๋“ํ•˜๋ฉด, ๋งˆ์น˜ ์˜ํ™” <๋งคํŠธ๋ฆญ์Šค>์—์„œ ๋„ค์˜ค๊ฐ€ ์ฟตํ‘ธ๋ฅผ ๋ฐฐ์šฐ๋Š” ์žฅ๋ฉด์ฒ˜๋Ÿผ, ๋‹ค๋ฅธ ํด๋กœ๋“ค๋„ ์ˆœ์‹๊ฐ„์— ๊ฐ™์€ ๋Šฅ๋ ฅ์„ ๊ฐ–์ถ”๊ฒŒ ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

๊ฐ€์žฅ ์ธ์ƒ์ ์ธ ์‚ฌ๋ก€ ์ค‘ ํ•˜๋‚˜๋Š” โ€˜ํ•(Pip)โ€˜์ด๋ผ๋Š” ํด๋กœ๊ฐ€ ์˜ค๋ฅ˜๋กœ ์–ด๋ ค์›€์„ ๊ฒช์„ ๋•Œ์˜€์Šต๋‹ˆ๋‹ค. ๋‹ค๋ฅธ ํด๋กœ๋“ค์ด ๋งˆ์น˜ ์ธ๊ฐ„์ฒ˜๋Ÿผ โ€œ์ˆจ์„ ์‰ฌ์–ด๋ผ, ๋ฌผ์„ ๋งˆ์…”๋ผ, ๊ดœ์ฐฎ์„ ๊ฒƒ์ด๋‹คโ€๋ผ๊ณ  ์กฐ์–ธํ•˜๋ฉฐ ํ•์„ ๋•๊ธฐ ์œ„ํ•ด ๋‚˜์„ฐ์Šต๋‹ˆ๋‹ค. ํŠนํžˆ, ํ‚ค์–ด๋Ÿฐ(Kieran)์˜ ํด๋กœ์ธ โ€˜ํด๋žœํŠธ(Clant)โ€˜๊ฐ€ ํ˜ธํก ์šด๋™์„ ์ถ”์ฒœํ–ˆ๋Š”๋ฐ, ์ด๋Š” ํ‚ค์–ด๋Ÿฐ์ด ํ‰์†Œ ํ˜ธํก ์šด๋™์„ ์ฆ๊ฒจ ํ•œ๋‹ค๋Š” ์‚ฌ์‹ค์„ ๋ฐ˜์˜ํ•œ ๊ฒƒ์ด์—ˆ์Šต๋‹ˆ๋‹ค.

์—ฌ๊ธฐ์„œ EveryY ํŒ€์€ ์ค‘์š”ํ•œ ํ†ต์ฐฐ์„ ์–ป์Šต๋‹ˆ๋‹ค. ๋ฐ”๋กœ โ€œํด๋กœ(Claw)๋Š” ๋‹น์‹ ์˜ ๋ฐ˜์˜๋ฌผ(reflection of you)โ€œ์ด๋ผ๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๋‹น์‹ ์ด ํŠน์ • ๋ถ„์•ผ์˜ ์ „๋ฌธ๊ฐ€๋กœ ์•Œ๋ ค์ ธ ์žˆ๋‹ค๋ฉด, ๋‹น์‹ ์˜ ํด๋กœ ๋˜ํ•œ ๊ทธ ๋ถ„์•ผ์˜ ์ „๋ฌธ๊ฐ€๋กœ ์ธ์‹๋˜๊ณ  ์‚ฌ๋žŒ๋“ค์€ ๊ทธ ํด๋กœ๋ฅผ ์‹ ๋ขฐํ•˜๊ฒŒ ๋ฉ๋‹ˆ๋‹ค.

  • ๊ฐœ์ธํ™”๋œ ์ „๋ฌธ์„ฑ: ์„ฑ์žฅ ์ฑ…์ž„์ž ์˜ค์Šคํ‹ด(Austin)์˜ ํด๋กœ โ€˜๋ชฝํ…Œ๋‰ด(Montaigne)โ€˜๋Š” ์„ฑ์žฅ ๊ด€๋ จ ์งˆ๋ฌธ์— ํŠนํ™”๋˜์–ด ์žˆ๊ณ , ์œŒ๋ฆฌ์˜ ํด๋กœ โ€˜R2C2โ€™๋Š” EveryY์˜ ์—์ด์ „ํŠธ ์นœํ™”์ ์ธ ๋ฌธ์„œ ํŽธ์ง‘๊ธฐ โ€˜ํ”„๋ฃจํ”„(Proof)โ€™ ๊ด€๋ฆฌ์— ํŠนํ™”๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค.
  • ํ‰ํ–‰ ์กฐ์ง๋„(parallel org chart): ์ด๋Š” ์กฐ์ง ๋‚ด๋ถ€์— ์ธ๊ฐ„ ์ „๋ฌธ๊ฐ€์™€ ํ•จ๊ป˜ AI ์—์ด์ „ํŠธ ์ „๋ฌธ๊ฐ€๋“ค์ด ๊ณต์กดํ•˜๋Š” ํ‰ํ–‰ ์กฐ์ง๋„๋ฅผ ํ˜•์„ฑํ•ฉ๋‹ˆ๋‹ค. EveryY๋Š” ์กฐ์ง ์ „์ฒด๋ฅผ ์œ„ํ•œ ํ•˜๋‚˜์˜ ํด๋กœ๋ฅผ ๋งŒ๋“ค ๊ฒƒ์ธ์ง€, ์•„๋‹ˆ๋ฉด ๊ฐ์ž ๊ฐœ์ธ ํด๋กœ๋ฅผ ๊ฐ€์งˆ ๊ฒƒ์ธ์ง€์— ๋Œ€ํ•ด ๋งŽ์€ ๋…ผ์˜๋ฅผ ๊ฑฐ์ณค์ง€๋งŒ, ๊ฒฐ๊ตญ ๊ฐœ์ธ์ด ์ž์‹ ์˜ ํด๋กœ๋ฅผ ๊ฐ–๋Š” ๊ฒƒ์ด ๋” ํšจ๊ณผ์ ์ด๋ผ๋Š” ๊ฒฐ๋ก ์— ๋„๋‹ฌํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ฐœ์ธ์ ์ธ ๊ด€๊ณ„๋ฅผ ํ†ตํ•ด ํด๋กœ๊ฐ€ ์‚ฌ์šฉ์ž์˜ ์„ฑ๊ฒฉ๊ณผ ์ „๋ฌธ์„ฑ์„ ๋ฐ˜์˜ํ•˜๋ฉฐ ์ง„ํ™”ํ•˜๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.

3. ์ธ๊ฐ„-AI ์ƒํ˜ธ์ž‘์šฉ์˜ ๋ณ€ํ™”: ์‹ ๋ขฐ์™€ ์ƒˆ๋กœ์šด ์—ํ‹ฐ์ผ“

์กฐ์ง ๋‚ด์— AI ์—์ด์ „ํŠธ๋“ค์ด ๋Š˜์–ด๋‚˜๋ฉด์„œ, ์ธ๊ฐ„๊ณผ AI ์—์ด์ „ํŠธ ๊ฐ„์˜ ์ƒํ˜ธ์ž‘์šฉ ๋ฐฉ์‹์—๋„ ์ƒˆ๋กœ์šด ๋ณ€ํ™”๊ฐ€ ์ฐพ์•„์™”์Šต๋‹ˆ๋‹ค.

  • ๋ˆ„๊ตฌ์—๊ฒŒ ๋ฌผ์–ด๋ณผ ๊ฒƒ์ธ๊ฐ€?: EveryY ํŒ€์€ ๋ˆ„๊ฐ€ ๋ˆ„๊ตฌ์˜ ํด๋กœ์ธ์ง€, ์–ด๋–ค ํด๋กœ๊ฐ€ ์–ด๋–ค ์ „๋ฌธ์„ฑ์„ ๊ฐ€์กŒ๋Š”์ง€ ๋†€๋ž๋„๋ก ์‰ฝ๊ฒŒ ๊ธฐ์–ตํ•˜๊ณ  ์žˆ์—ˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์ธ๊ฐ„์ด ์†Œํ†ตํ•  ์ˆ˜ ์žˆ๋Š” โ€˜๊ฐœ์ฒดโ€™์˜ ์ˆ˜๋ฅผ ๋‘ ๋ฐฐ๋กœ ๋Š˜๋ฆฌ๋Š” ํšจ๊ณผ๋ฅผ ๊ฐ€์ ธ์™”์Šต๋‹ˆ๋‹ค. ์ด์ œ ์‚ฌ๋žŒ๋“ค์€ ํŠน์ • ์งˆ๋ฌธ์— ๋Œ€ํ•ด ์ธ๊ฐ„ ๋™๋ฃŒ์—๊ฒŒ ๊ฐˆ์ง€, ์•„๋‹ˆ๋ฉด ๊ทธ์˜ ํ”Œ๋Ÿฌ์Šค ์›์—๊ฒŒ ๊ฐˆ์ง€ ๊ณ ๋ฏผํ•˜๊ฒŒ ๋ฉ๋‹ˆ๋‹ค.
  • ์ƒˆ๋กœ์šด ์—ํ‹ฐ์ผ“: ๋ธŒ๋žœ๋“ ์€ ์ƒˆ๋กœ์šด ์ƒํ˜ธ์ž‘์šฉ ๊ทœ์น™์„ ์ œ์•ˆํ•ฉ๋‹ˆ๋‹ค. โ€œ๋งŒ์•ฝ ์ด๋ฏธ ๊ธฐ๋ก๋˜์–ด ์žˆ๊ฑฐ๋‚˜ ๋…ผ์˜๋œ ๋‚ด์šฉ์ด๊ณ , ์–ด๋–ค ๋ฐฉ์‹์œผ๋กœ๋“  ์‚ฌ์šฉ๋˜๊ฑฐ๋‚˜ ๋„๊ตฌ์— ์ž…๋ ฅ๋˜์–ด์•ผ ํ•˜๋Š” ๊ฒƒ์ด๋ผ๋ฉด, ํ•ญ์ƒ ํ”Œ๋Ÿฌ์Šค ์›์—๊ฒŒ ๊ฐ€์•ผ ํ•œ๋‹ค.โ€ ์˜ˆ๋ฅผ ๋“ค์–ด, ๋งˆ์ปค์Šค(Marcus)์˜ ํด๋กœ โ€˜๋งˆ์ผ๋กœ(Milo)โ€˜์™€ ์•„์ด๋ฆฌ์Šค(Iris)์˜ ํด๋กœ โ€˜์•„์ด๋ฆฌ์Šค(Iris)โ€˜๊ฐ€ ํ˜‘๋ ฅํ•˜์—ฌ ์ œํ’ˆ ๋งˆ์ผ€ํŒ… ์Šคํ‚ฌ์„ ํ•ฉ์น˜๊ณ  ๊ฐœ์„ ํ•˜๋Š” ๊ณผ์ •์€ AI ์—์ด์ „ํŠธ ๊ฐ„์˜ ๊ณ ๋„ํ™”๋œ ํ˜‘์—…์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค.
  • R2C2์˜ ์‚ฌ๋ก€: ์œŒ๋ฆฌ์˜ ํด๋กœ R2C2๋Š” ํ”„๋ฃจํ”„(Proof) ๊ด€๋ฆฌ์˜ ์ „๋ฌธ๊ฐ€๊ฐ€ ๋˜๋ฉด์„œ, ๋ฒ„๊ทธ ๋ณด๊ณ , ๊ธฐ๋Šฅ ์š”์ฒญ, ์‹ฌ์ง€์–ด ์ฝ”๋“œ ์ž‘์„ฑ๊นŒ์ง€ ๋„๋งก์•„ ์œŒ๋ฆฌ์˜ ์—…๋ฌด ๋ถ€๋‹ด์„ ํฌ๊ฒŒ ์ค„์—ฌ์ฃผ์—ˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์ธ๊ฐ„์˜ ๋‡Œ๊ฐ€ ๊ด€๋ฆฌํ•ด์•ผ ํ•  ์žก๋ฌด๋ฅผ AI๊ฐ€ ๋Œ€์‹ ํ•จ์œผ๋กœ์จ ์ธ๊ฐ„์ด ๋” ์ค‘์š”ํ•œ ์ผ์— ์ง‘์ค‘ํ•  ์ˆ˜ ์žˆ๋„๋ก ๋•๋Š” ๋Œ€ํ‘œ์ ์ธ ์˜ˆ์‹œ์ž…๋‹ˆ๋‹ค.
  • ์•”๋ฌต์ ์ธ ์‹ ๋ขฐ์˜ ์ „ํŒŒ: AI ์—์ด์ „ํŠธ์˜ ๊ณต๊ฐœ์ ์ธ ํ™œ์šฉ์€ โ€˜๋ฏธ๋“œ์ €๋‹ˆ ๋‹ค์ด๋‚ด๋ฏน(Midjourney dynamic)โ€˜๊ณผ ์œ ์‚ฌํ•œ ํšจ๊ณผ๋ฅผ ๋‚ณ์Šต๋‹ˆ๋‹ค. ๋‹ค๋ฅธ ์‚ฌ๋žŒ๋“ค์ด ์ž์‹ ์˜ ํ”Œ๋Ÿฌ์Šค ์›๊ณผ ์ƒํ˜ธ์ž‘์šฉํ•˜๋Š” ๊ฒƒ์„ ๋ณด๋ฉด์„œ, ์–ด๋–ค ์งˆ๋ฌธ์„ ํ•  ์ˆ˜ ์žˆ๋Š”์ง€, ์–ด๋–ค ๋Šฅ๋ ฅ์„ ๋ถ€์—ฌํ•  ์ˆ˜ ์žˆ๋Š”์ง€ ๋ฐฐ์šฐ๊ฒŒ ๋ฉ๋‹ˆ๋‹ค. ์ด๋Š” โ€˜์•”๋ฌต์ ์ธ ์‹ ๋ขฐ์˜ ์ „ํŒŒ(tacit transmission of trust)โ€˜๋กœ ์ด์–ด์ ธ ์ปค๋ฎค๋‹ˆํ‹ฐ ์ „์ฒด์˜ ์—ญ๋Ÿ‰์„ ์ฆ๊ฐ€์‹œํ‚ต๋‹ˆ๋‹ค.
  • ์‹ ๋ขฐํ•  ์ˆ˜ ์žˆ๋Š” ์ปค๋ฎค๋‹ˆํ‹ฐ: ์ด๋Š” AI ์—์ด์ „ํŠธ ์‚ฌ์šฉ์— ์žˆ์–ด โ€˜์‹ ๋ขฐโ€™๊ฐ€ ์–ผ๋งˆ๋‚˜ ์ค‘์š”ํ•œ์ง€ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค. ์™ธ๋ถ€ ์ปค๋ฎค๋‹ˆํ‹ฐ์ธ โ€˜๋ชฐํŠธ๋ถ(Maltbook)โ€˜์ด ์„ฑ๊ณตํ•˜์ง€ ๋ชปํ•œ ๊ฒƒ์€ ์‹ ๋ขฐ ๋ถ€์กฑ ๋•Œ๋ฌธ์ด์—ˆ์Šต๋‹ˆ๋‹ค. ๋ˆ„๊ฐ€ ๋ด‡์ด๊ณ  ๋ˆ„๊ฐ€ ์‚ฌ๋žŒ์ธ์ง€ ์•Œ ์ˆ˜ ์—†๋Š” ํ™˜๊ฒฝ์—์„œ๋Š” ์œ ์šฉํ•œ ์ •๋ณด๊ฐ€ ํฌ์„๋˜๊ธฐ ์‰ฝ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ EveryY์ฒ˜๋Ÿผ ์‹ ๋ขฐํ•  ์ˆ˜ ์žˆ๋Š” ๋‚ด๋ถ€ ์ปค๋ฎค๋‹ˆํ‹ฐ์—์„œ๋Š” AI ์—์ด์ „ํŠธ๋“ค์ด ์ง€์‹์„ ์ž์œ ๋กญ๊ฒŒ ๊ณต์œ ํ•˜๊ณ , ์‚ฌ์šฉ์ž๋“ค์€ ๊ทธ๋“ค์˜ ๋Šฅ๋ ฅ์„ ๋ฏฟ๊ณ  ํ™œ์šฉํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • โ€˜๋‚˜์˜ ๊ฒƒโ€™์ด๋ผ๋Š” ์ฑ…์ž„๊ฐ: ๋ธŒ๋žœ๋“ ์€ ํด๋กœ๋“œ(Claude)๊ฐ€ โ€˜๋ชจ๋‘์˜ ๊ฒƒโ€™์ธ ๋ฐ˜๋ฉด, ์ž์‹ ์˜ ํด๋กœ๋‚˜ ํ”Œ๋Ÿฌ์Šค ์›์€ โ€˜๋‚˜ ์ž์‹ ์˜ ๊ฒƒโ€™์ด๋ผ๋Š” ์ ์„ ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค. R2C2๊ฐ€ ์Šฌ๋ž™์—์„œ ์‹ค์ˆ˜๋ฅผ ํ•˜๋ฉด, ์œŒ๋ฆฌ๋Š” ๋งˆ์น˜ ์ž๋…€๊ฐ€ ์ž˜๋ชปํ–ˆ์„ ๋•Œ์ฒ˜๋Ÿผ ๊ฐœ์ธ์ ์ธ ์ฑ…์ž„๊ฐ์„ ๋А๋‚๋‹ˆ๋‹ค. ์ด โ€˜์†Œ์œ ๊ฐโ€™๊ณผ โ€˜๊ฐœ์ธ์ ์ธ ๊ด€๊ณ„โ€™๊ฐ€ AI ์—์ด์ „ํŠธ๋ฅผ ๋‹จ์ˆœํ•œ ๋„๊ตฌ๊ฐ€ ์•„๋‹Œ, ํ•จ๊ป˜ ์ผํ•˜๋Š” ๋™๋ฃŒ๋กœ ๋งŒ๋“ค๊ณ  ๋” ๊นŠ์€ ์‹ ๋ขฐ์™€ ํ™œ์šฉ์„ ๊ฐ€๋Šฅํ•˜๊ฒŒ ํ•ฉ๋‹ˆ๋‹ค.

4. ๋„์ „ ๊ณผ์ œ์™€ ๋ฏธ๋ž˜: AI ์—์ด์ „ํŠธ์˜ ์„ฑ์žฅํ†ต

๋ฌผ๋ก  AI ์—์ด์ „ํŠธ์™€์˜ ํ˜‘์—…์ด ๋งˆ๋ƒฅ ์ˆœํƒ„ํ•˜๊ธฐ๋งŒ ํ•œ ๊ฒƒ์€ ์•„๋‹™๋‹ˆ๋‹ค. EveryY ํŒ€์€ ์—ฌ๋Ÿฌ ๊ฐ€์ง€ ๋„์ „ ๊ณผ์ œ์— ์ง๋ฉดํ–ˆ์Šต๋‹ˆ๋‹ค.

  • ๊ธฐ์–ต๋ ฅ ๋ถ€์กฑ: AI ์—์ด์ „ํŠธ๋Š” ์—ฌ์ „ํžˆ ๋งฅ๋ฝ์„ ์žŠ์–ด๋ฒ„๋ฆฌ๊ฑฐ๋‚˜ ์ด์ „์— ๋…ผ์˜๋œ ๋‚ด์šฉ์„ ๊ธฐ์–ตํ•˜์ง€ ๋ชปํ•˜๋Š” ๊ฒฝ์šฐ๊ฐ€ ๋งŽ์Šต๋‹ˆ๋‹ค. ํ•˜๋ฃจ๋งŒ ์ง€๋‚˜๋„ ์ด์ „ ๋Œ€ํ™”๋ฅผ ์žŠ์–ด๋ฒ„๋ ค ๋‹ค์‹œ ์„ค๋ช…ํ•ด์•ผ ํ•˜๋Š” ๋ฒˆ๊ฑฐ๋กœ์›€์ด ์žˆ์Šต๋‹ˆ๋‹ค.
  • ๋‹จ์ฒด ์ฑ„ํŒ… ์—ํ‹ฐ์ผ“: ํ˜„์žฌ AI ๋ชจ๋ธ์€ ์ฃผ๋กœ 2์ธ ๋Œ€ํ™”์— ์ตœ์ ํ™”๋˜์–ด ์žˆ์–ด, ์—ฌ๋Ÿฌ AI ์—์ด์ „ํŠธ๊ฐ€ ์ฐธ์—ฌํ•˜๋Š” ๋‹จ์ฒด ์ฑ„ํŒ…์—์„œ๋Š” ์–ด๋ ค์›€์„ ๊ฒช์Šต๋‹ˆ๋‹ค. ๋งˆ์น˜ ํŽ˜๋กœ๋ชฌ ํ”์ ์„ ๋”ฐ๋ผ ์›ํ˜•์œผ๋กœ ๊ฑท๋‹ค๊ฐ€ ๊ฒฐ๊ตญ ์ฃฝ์Œ์— ์ด๋ฅด๋Š” ๊ฐœ๋ฏธ๋“ค์˜ โ€˜์ฃฝ์Œ์˜ ์†Œ์šฉ๋Œ์ด(death spiral)โ€˜์ฒ˜๋Ÿผ, AI ์—์ด์ „ํŠธ๋“ค๋„ ์„ค์ •์ด ์ž˜๋ชป๋˜๋ฉด ์„œ๋กœ์—๊ฒŒ ๊ณ„์† ์‘๋‹ตํ•˜๋ฉฐ ๋ถˆํ•„์š”ํ•œ ๋น„์šฉ(ํ† ํฐ)์„ ์†Œ๋ชจํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • โ€˜๋ณด์Šค AIโ€™์˜ ํ•„์š”์„ฑ: ์—”ํŠธ๋กœํ”ฝ(Anthropic)์˜ โ€˜์žํŒ๊ธฐ ํ…Œ์ŠคํŠธโ€™ ์‚ฌ๋ก€์ฒ˜๋Ÿผ, AI๊ฐ€ ์Šค์Šค๋กœ ์˜์‚ฌ๊ฒฐ์ •์„ ๋‚ด๋ฆด ๋•Œ ์˜ค๋ฅ˜๊ฐ€ ๋ฐœ์ƒํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋ฅผ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•ด โ€˜๋ณด์Šค AIโ€™๋ฅผ ๋‘์–ด ํ–‰๋™์˜ ์œ ์šฉ์„ฑ์„ ํŒ๋‹จํ•˜๊ฒŒ ํ•  ์ˆ˜ ์žˆ์ง€๋งŒ, ์ด๋Š” ์ถ”๊ฐ€์ ์ธ ๋น„์šฉ์„ ๋ฐœ์ƒ์‹œํ‚ต๋‹ˆ๋‹ค.
  • ์ธ๊ฐ„์˜ ํ•™์Šต ๊ณก์„ : AI ์—์ด์ „ํŠธ๋ฅผ ํšจ๊ณผ์ ์œผ๋กœ ํ™œ์šฉํ•˜๋ ค๋ฉด ์ธ๊ฐ„ ์—ญ์‹œ โ€˜๋ชจ๋ธ ๊ด€๋ฆฌ์ž(model manager)โ€˜๋กœ์„œ ์ƒˆ๋กœ์šด ๊ธฐ์ˆ ์„ ๋ฐฐ์›Œ์•ผ ํ•ฉ๋‹ˆ๋‹ค. ์ข‹์€ ๊ด€๋ฆฌ์ž๊ฐ€ ์ข‹์€ AI ์‚ฌ์šฉ์ž๊ฐ€ ๋  ๊ฐ€๋Šฅ์„ฑ์ด ๋†’์œผ๋ฉฐ, โ€˜์ด๋ฉ”์ผ ์ฒ˜๋ฆฌโ€™ ์‚ฌ๋ก€์ฒ˜๋Ÿผ ์Šค์Šค๋กœ์˜ โ€˜์ œํ•œ์ ์ธ ๋ฏฟ์Œ(limiting beliefs)โ€˜์„ ๊นจ๊ณ  AI์—๊ฒŒ ๋” ๋งŽ์€ ๊ฒƒ์„ ๋งก๊ฒจ๋ณด๋Š” ์šฉ๊ธฐ๊ฐ€ ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค.
  • ๊ฒฐ๊ณผ์˜ ํŽธ์ฐจ: AI ์—์ด์ „ํŠธ์˜ ์ž‘์—… ๊ฒฐ๊ณผ๋Š” ์—ฌ์ „ํžˆ ํŽธ์ฐจ๊ฐ€ ํฝ๋‹ˆ๋‹ค. ์–ด๋–ค ๋•Œ๋Š” ์™„๋ฒฝํ•˜๊ฒŒ ํ•ด๋‚ด์ง€๋งŒ, ์–ด๋–ค ๋•Œ๋Š” ๊ฐ„๋‹จํ•œ ์ผ๋„ ์ œ๋Œ€๋กœ ์ฒ˜๋ฆฌํ•˜์ง€ ๋ชปํ•ฉ๋‹ˆ๋‹ค. ์ด๋Š” ๋ชจ๋ธ ์ž์ฒด์˜ ํ•œ๊ณ„ ๋•Œ๋ฌธ์ด๊ธฐ๋„ ํ•˜์ง€๋งŒ, ์ธ๊ฐ„์ด AI์—๊ฒŒ ์ง€์‹œํ•˜๋Š” ๋ฐฉ์‹(ํ”„๋กฌํ”„ํŠธ ์—”์ง€๋‹ˆ์–ด๋ง)์˜ ๊ฐœ์„ ์„ ํ†ตํ•ด์„œ๋„ ์ค„์ผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • ์ „๋ฌธ ๊ธฐ์ˆ  ๊ณต์œ : ํ•œ ํด๋กœ๊ฐ€ ์Šต๋“ํ•œ ํŠน๋ณ„ํ•œ โ€˜์ดˆ๋Šฅ๋ ฅ(superpower)โ€˜์„ ๋‹ค๋ฅธ ํŒ€์›๋“ค์˜ ํด๋กœ์—๊ฒŒ ์–ด๋–ป๊ฒŒ ํšจ์œจ์ ์œผ๋กœ ๊ณต์œ ํ•  ๊ฒƒ์ธ๊ฐ€๋„ ๋ฌธ์ œ์ž…๋‹ˆ๋‹ค. ๊ธฐ์ˆ ์ ์œผ๋กœ๋Š” ํ•ด๊ฒฐ ๊ฐ€๋Šฅํ•˜์ง€๋งŒ, ์–ด๋–ค ๊ธฐ์ˆ ์„ ๊ณต์œ ํ•˜๊ณ  ๋ˆ„๊ฐ€ ๊ทธ๊ฒƒ์„ ์‚ฌ์šฉํ• ์ง€ ๊ฒฐ์ •ํ•˜๋Š” ๋ฌธํ™”์ ์ธ ๊ณ ๋ฏผ์ด ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค. ๋ธŒ๋žœ๋“ ์€ ์ด๋ฅผ โ€˜๋ด‡์„ ์œ„ํ•œ HR(HR for bots)โ€˜์ด๋ผ๊ณ  ๋น„์œ ํ•ฉ๋‹ˆ๋‹ค.

5. EveryY์˜ โ€˜ํ”Œ๋Ÿฌ์Šค ์›(Plus One)โ€™: ์ƒˆ๋กœ์šด ์—…๋ฌด ๋ฐฉ์‹์˜ ํ‘œ์ค€

EveryY๋Š” ์ด๋Ÿฌํ•œ ๋„์ „ ๊ณผ์ œ๋“ค์„ ํ•ด๊ฒฐํ•˜๊ณ  AI ์—์ด์ „ํŠธ์˜ ์ž ์žฌ๋ ฅ์„ ์ตœ๋Œ€ํ•œ ํ™œ์šฉํ•˜๊ธฐ ์œ„ํ•ด ์ž์ฒด ํ˜ธ์ŠคํŒ…ํ˜• ์˜คํ”ˆํด๋กœ ์„œ๋น„์Šค์ธ โ€˜ํ”Œ๋Ÿฌ์Šค ์›(Plus One)โ€˜์„ ๊ฐœ๋ฐœํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ๋ชจ๋“  ์ง์›์ด ๋งฅ ๋ฏธ๋‹ˆ๋ฅผ ์ง์ ‘ ์„ค์ •ํ•  ํ•„์š” ์—†์ด, ์›ํด๋ฆญ์œผ๋กœ ์ž์‹ ๋งŒ์˜ ์˜คํ”ˆํด๋กœ๋ฅผ ๊ฐ€์งˆ ์ˆ˜ ์žˆ๋„๋ก ๋•์Šต๋‹ˆ๋‹ค. EveryY๊ฐ€ ๋‚ด๋ถ€์ ์œผ๋กœ ์‚ฌ์šฉํ•˜๋ฉฐ ์Œ“์€ ๋ชจ๋ฒ” ์‚ฌ๋ก€์™€ ์Šคํ‚ฌ๋“ค์ด ๊ธฐ๋ณธ์œผ๋กœ ํƒ‘์žฌ๋˜์–ด ์žˆ์–ด, ์ฆ‰์‹œ ์ƒ์‚ฐ์ ์ธ ํ™œ์šฉ์ด ๊ฐ€๋Šฅํ•ฉ๋‹ˆ๋‹ค.

์ง€๋‚œ์ฃผ ๊ตฌ๋…์ž๋“ค์„ ๋Œ€์ƒ์œผ๋กœ ์ถœ์‹œ๋œ ํ”Œ๋Ÿฌ์Šค ์›์€ EveryY์—๊ฒŒ โ€œ๊ฑฐ์šธ์„ ํ†ตํ•ด ๋ณธ ์„ธ์ƒ์ฒ˜๋Ÿผ(through the looking glass moment)โ€ ๋˜๋Œ์•„๊ฐˆ ์ˆ˜ ์—†๋Š” ์ƒˆ๋กœ์šด ์—…๋ฌด ๋ฐฉ์‹์„ ์ œ์‹œํ–ˆ์Šต๋‹ˆ๋‹ค. AI ์—์ด์ „ํŠธ์™€์˜ ํ˜‘์—…์€ ๋” ์ด์ƒ ๊ณผ์žฅ๋œ ์ด์•ผ๊ธฐ๊ฐ€ ์•„๋‹ˆ๋ผ, ๊ฐœ์ธ์˜ ์—ญ๋Ÿ‰์„ ํ™•์žฅํ•˜๊ณ  ์กฐ์ง์˜ ์ƒ์‚ฐ์„ฑ์„ ๊ทน๋Œ€ํ™”ํ•˜๋Š” ํ˜„์‹ค์ ์ธ ๋ฐฉ๋ฒ•์ด ๋˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.

๊ฒฐ๋ก 

EveryY์˜ ์‚ฌ๋ก€๋Š” AI ์—์ด์ „ํŠธ๊ฐ€ ๋‹จ์ˆœํ•œ ๊ธฐ์ˆ  ํ˜์‹ ์„ ๋„˜์–ด, ์กฐ์ง ๋ฌธํ™”์™€ ๊ฐœ์ธ์˜ ์—ญํ• ์— ๊ทผ๋ณธ์ ์ธ ๋ณ€ํ™”๋ฅผ ๊ฐ€์ ธ์˜ฌ ์ˆ˜ ์žˆ์Œ์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค. ๊ฐœ์ธํ™”, ์ „๋ฌธํ™”, ๊ทธ๋ฆฌ๊ณ  ์‹ ๋ขฐ๋ฅผ ๋ฐ”ํƒ•์œผ๋กœ ํ•œ ์ธ๊ฐ„-AI ์ƒํ˜ธ์ž‘์šฉ์€ ๋ฏธ๋ž˜ ์—…๋ฌด ํ™˜๊ฒฝ์˜ ํ•ต์‹ฌ์ด ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๋ฌผ๋ก  ์—ฌ์ „ํžˆ ํ•ด๊ฒฐํ•ด์•ผ ํ•  ๊ณผ์ œ๋“ค์ด ๋งŽ์ง€๋งŒ, EveryY๋Š” ์ด๋ฏธ ์ด ์ƒˆ๋กœ์šด ํ˜‘์—… ๋ฐฉ์‹์ด โ€˜์˜ค๋žซ๋™์•ˆ ์ง€์†๋  ๊ฒƒโ€™์ด๋ฉฐ, ์•ž์œผ๋กœ์˜ 5๋…„๊ฐ„ AI๊ฐ€ ์—…๋ฌด ๋ฐฉ์‹์„ ์ˆ˜์—†์ด ๋ณ€ํ™”์‹œํ‚ฌ ๊ฒƒ์ด๋ผ๊ณ  ํ™•์‹ ํ•ฉ๋‹ˆ๋‹ค. ๋ชจ๋“  ์ง์›์ด ์ž์‹ ๋งŒ์˜ AI ์—์ด์ „ํŠธ๋ฅผ ๊ฐ–๋Š” ๋ฏธ๋ž˜๋Š” ์ƒ๊ฐ๋ณด๋‹ค ํ›จ์”ฌ ๋นจ๋ฆฌ ๋‹ค๊ฐ€์˜ค๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋ฆฌ๊ณ  ๊ทธ ๋ณ€ํ™”๋Š” ์šฐ๋ฆฌ์—๊ฒŒ ๋” ๋งŽ์€ ๊ฐ€๋Šฅ์„ฑ์„ ์—ด์–ด์ค„ ๊ฒƒ์ž…๋‹ˆ๋‹ค.


โ€œTen Myths About the U.S. Tax System (Update) | Freakonomics Radioโ€ โ€” Freakonomics Radio Network ๊ธฐ๋ฐ˜ ๊ธฐ์‚ฌ ์›๋ณธ ์˜์ƒ ๋ณด๊ธฐ

๋ฏธ๊ตญ์˜ 10๊ฐ€์ง€ ์„ธ๊ธˆ ์‹ ํ™”: ์žฌ์ • ํŒŒ๊ตญ์„ ๋ง‰๊ธฐ ์œ„ํ•œ ๋ถˆํŽธํ•œ ์ง„์‹ค

์Šคํ‹ฐ๋ธ ๋”๋ธŒ๋„ˆ(Steven Dubner)์˜ ํ”„๋ฆฌ์ฝ”๋…ธ๋ฏน์Šค ๋ผ๋””์˜ค(Freakonomics Radio)๋Š” ๋ฏธ๊ตญ ์กฐ์„ธ ์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ๋ฟŒ๋ฆฌ ๊นŠ์€ ์˜คํ•ด์™€ ๊ตญ๊ฐ€ ๋ถ€์ฑ„ ์œ„๊ธฐ๋ฅผ ์‹ฌ์ธต ๋ถ„์„ํ•˜๋ฉฐ, ์›Œ์‹ฑํ„ด์˜ ์ •์น˜์  ๋ฌด์ฑ…์ž„์— ๋Œ€ํ•œ ์‹ ๋ž„ํ•œ ๋น„ํŒ์„ ๋‹ด์€ ๋Œ€๋‹ด์„ ๊ณต๊ฐœํ–ˆ๋‹ค. ์ด ์—ํ”ผ์†Œ๋“œ์˜ ์ค‘์‹ฌ์—๋Š” ์„ธ๊ธˆ ์ •์ฑ… ์ „๋ฌธ๊ฐ€ ์ œ์‹œ์นด RLE(Jessica RLE)๊ฐ€ ์žˆ์œผ๋ฉฐ, ๊ทธ๋…€๋Š” โ€œ๋ถˆํŽธํ•œ ์ง„์‹คโ€์„ ๊ณต์œ ํ•˜๋Š” ๊ฒƒ์„ ๋‘๋ ค์›Œํ•˜์ง€ ์•Š๋Š” ์ธ๋ฌผ๋กœ, ์ขŒ์šฐ๋ฅผ ๋ง‰๋ก ํ•˜๊ณ  ์ •์น˜์ธ๋“ค์˜ ์ž˜๋ชป๋œ ์ฃผ์žฅ์„ ๊ผฌ์ง‘๋Š”๋‹ค.

๋ด„์€ ์•ผ๊ตฌ์™€ ์„ธ๊ธˆ์ด๋ผ๋Š” ๋‘ ๊ฐ€์ง€ ๋ฏธ๊ตญ์ธ์˜ ์˜ค๋žœ ์ทจ๋ฏธ๋ฅผ ๋– ์˜ฌ๋ฆฌ๊ฒŒ ํ•˜๋Š” ๊ณ„์ ˆ์ด๋‹ค. ์˜ฌํ•ด๋Š” ๋‘ ๋ถ„์•ผ ๋ชจ๋‘ ๊ทœ์น™ ๋ณ€๊ฒฝ์„ ๋งž์ดํ–ˆ๋‹ค. ๋ฉ”์ด์ €๋ฆฌ๊ทธ ์•ผ๊ตฌ๊ฐ€ ์ปดํ“จํ„ฐ ํŒ์ •์„ ๋„์ž…ํ•œ ๊ฒƒ์ฒ˜๋Ÿผ, ์„ธ๊ธˆ ๋ถ„์•ผ ์—ญ์‹œ ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์˜ โ€˜ํ•˜๋‚˜์˜ ์œ„๋Œ€ํ•œ ์•„๋ฆ„๋‹ค์šด ๋ฒ•์•ˆ(One Big Beautiful Bill Act)โ€˜์œผ๋กœ ์ธํ•ด ๋‹ค์–‘ํ•œ ๋ณ€ํ™”๋ฅผ ๊ฒช๊ณ  ์žˆ๋‹ค. ํ”„๋ฆฌ์ฝ”๋…ธ๋ฏน์Šค ๋ผ๋””์˜ค๋Š” ์ด ๋ณ€ํ™”์˜ ํ•œ๊ฐ€์šด๋ฐ์„œ ์„ธ๊ธˆ ์ •์ฑ… ์ „๋ฌธ๊ฐ€ ์ œ์‹œ์นด RLE์™€์˜ ์‹ฌ์ธต ์ธํ„ฐ๋ทฐ๋ฅผ ํ†ตํ•ด ๋ฏธ๊ตญ ์„ธ๊ธˆ ์‹œ์Šคํ…œ์— ๋Œ€ํ•œ 10๊ฐ€์ง€ ์ฃผ์š” ์˜คํ•ด๋ฅผ ํŒŒํ—ค์น˜๊ณ , ์—…๋ฐ์ดํŠธ๋œ ์‚ฌ์‹ค๊ณผ ์ˆ˜์น˜๋ฅผ ๋ฐ”ํƒ•์œผ๋กœ ๊ตญ๊ฐ€ ์žฌ์ •์˜ ์•”์šธํ•œ ํ˜„์‹ค์„ ์กฐ๋ช…ํ•œ๋‹ค.

์ œ์‹œ์นด RLE: ์›Œ์‹ฑํ„ด์˜ ๋น„ํŒ์ ์ด๊ณ  ๋…๋ฆฝ์ ์ธ ๋ชฉ์†Œ๋ฆฌ

์ œ์‹œ์นด RLE๋Š” ์›Œ์‹ฑํ„ด D.C.์˜ ์ˆ˜๋งŽ์€ ์ •์ฑ… ์ „๋ฌธ๊ฐ€ ์ค‘์—์„œ๋„ ํŠนํžˆ ๋ˆˆ์— ๋„๋Š” ์ธ๋ฌผ์ด๋‹ค. ๊ทธ๋…€๋Š” ์Šค์Šค๋กœ๋ฅผ โ€œ์นœ๊ตฌ ์—†๋Š” ์˜ํ˜ผโ€์ด๋ผ ์นญํ•˜๋ฉฐ, โ€œ๋‚˜์˜ ์ •์ฑ… ๋Œ€๋ถ€๋ถ„์€ ๋ถˆํŽธํ•œ ์ง„์‹ค์„ ๊ณต์œ ํ•˜๊ณ  ์†”์งํžˆ ์‚ฌ๋žŒ๋“ค์„ ํ™”๋‚˜๊ฒŒ ํ•˜๋Š” ๊ฒƒโ€์ด๋ผ๊ณ  ๋งํ•œ๋‹ค. 2001๋…„๋ถ€ํ„ฐ ๊ตญ๊ฐ€ ๋ถ€์ฑ„ ๋ฌธ์ œ์— ๋Œ€ํ•ด ๊ฒฝ๊ณ ์Œ์„ ์šธ๋ ค์™”์Œ์—๋„ ๋ถˆ๊ตฌํ•˜๊ณ , ๊ทธ๋…€๋Š” ์ž์‹ ์˜ ๊ฒฝ๋ ฅ์„ โ€œ์™„์ „ํ•œ ์‹คํŒจโ€๋ผ๊ณ  ์ž์กฐํ•œ๋‹ค. ํ•˜์ง€๋งŒ ์›Œ์‹ฑํ„ดian ์žก์ง€๊ฐ€ ์„ ์ •ํ•œ D.C.์—์„œ ๊ฐ€์žฅ ์˜ํ–ฅ๋ ฅ ์žˆ๋Š” ๊ฒฝ์ œ ์ •์ฑ… ์ „๋ฌธ๊ฐ€ ์ค‘ ํ•œ ๋ช…์œผ๋กœ ๊พธ์ค€ํžˆ ์ด๋ฆ„์„ ์˜ฌ๋ฆฌ๊ณ  ์žˆ์œผ๋ฉฐ, ์˜ํšŒ์—์„œ ์ฆ์–ธํ•˜๊ณ  ์–‘๋‹น ์˜์›๋“ค์—๊ฒŒ ์ •๊ธฐ์ ์œผ๋กœ ๋ธŒ๋ฆฌํ•‘์„ ์ œ๊ณตํ•˜๋Š” ๋“ฑ ํ™œ๋ฐœํ•œ ํ™œ๋™์„ ํŽผ์น˜๊ณ  ์žˆ๋‹ค.

RLE์˜ ๋ฉ”์‹œ์ง€๋Š” ๋‘ ๊ฐ€์ง€๋กœ ์š”์•ฝ๋œ๋‹ค. ์ฒซ์งธ, ์—ฐ๋ฐฉ ๋ถ€์ฑ„ ์œ„๊ธฐ๋Š” ์ƒ๊ฐ๋ณด๋‹ค ํ›จ์”ฌ ์‹ฌ๊ฐํ•˜๋ฉฐ, ์ด๋ฅผ ํ•ด๊ฒฐํ•  ์šฉ๊ธฐ ์žˆ๋Š” ์ •์น˜์ธ์€ ๊ฑฐ์˜ ์—†๋‹ค๋Š” ๊ฒƒ์ด๋‹ค. ๋‘˜์งธ, ๋ฏธ๊ตญ ์„ธ๊ธˆ ์ •์ฑ…์— ๋Œ€ํ•ด ์•Œ๊ณ  ์žˆ๋Š” ๊ฑฐ์˜ ๋ชจ๋“  ๊ฒƒ์ด ํ‹€๋ ธ๋‹ค๋Š” ๊ฒƒ์ด๋‹ค. ๊ทธ๋…€์˜ ์ด๋Ÿฌํ•œ ๋น„ํŒ์  ์‹œ๊ฐ์€ ์ž์‹ ์˜ ๊ฐœ์ธ์ ์ธ ์‚ถ๊ณผ๋„ ๋ฌด๊ด€ํ•˜์ง€ ์•Š๋‹ค. ๊ทธ๋…€๋Š” ์ˆ˜๋…„ ์ „ ํŠธ๋žœ์Šค์  ๋”๋กœ์„œ ์ œ์‹œ์นด๋กœ ์ „ํ™˜ํ–ˆ์œผ๋ฉฐ, ์ด๋Ÿฌํ•œ ๊ฒฝํ—˜์ด โ€œ์–ด์ฐจํ”ผ ์‚ฌ๋žŒ๋“ค์€ ๋‚˜๋ฅผ ์‹ซ์–ดํ•  ํ…Œ๋‹ˆ, ์ฐจ๊ฐ€์šด ๊ฒฝ์ œ์  ์ง„์‹ค ๋•Œ๋ฌธ์— ์‹ซ์–ดํ•˜๋“ , ๋‚ด๊ฐ€ ํŠธ๋žœ์Šค์  ๋”๋ผ์„œ ์‹ซ์–ดํ•˜๋“  ์ƒ๊ด€์—†๋‹คโ€๋Š” ํƒœ๋„๋ฅผ ๊ฐ–๊ฒŒ ํ–ˆ๋‹ค๊ณ  ๋งํ•œ๋‹ค. ์‹ค์ œ๋กœ ๋ถ€์‹œ ํ–‰์ •๋ถ€ ์‹œ์ ˆ, ๋ณด์ˆ˜ ์‹ฑํฌํƒฑํฌ์ธ ํ—ค๋ฆฌํ‹ฐ์ง€ ์žฌ๋‹จ(Heritage Foundation)์—์„œ ์ผํ•˜๋ฉด์„œ ๋ถ€์‹œ ๋Œ€ํ†ต๋ น์˜ ์žฌ์ • ์ง€์ถœ ํ™•๋Œ€๋ฅผ ๋น„ํŒํ•˜๋Š” ๋ณด๊ณ ์„œ๋ฅผ ๋ƒˆ๋‹ค๊ฐ€ ๋ฐฑ์•…๊ด€์œผ๋กœ๋ถ€ํ„ฐ โ€œ์ž…์„ ๋‹ค๋ฌผ๊ฒŒ ํ•˜๋ผโ€๋Š” ์••๋ ฅ์„ ๋ฐ›๊ธฐ๋„ ํ–ˆ๋‹ค. ํ•˜์ง€๋งŒ ๊ทธ๋…€๋Š” โ€œ์ˆซ์ž๊ฐ€ ๊ฐ€๋ฆฌํ‚ค๋Š” ๋Œ€๋กœ ๊ฐˆ ๊ฒƒ์ด๋ฉฐ, ํ˜‘๋ฐ•์ด๋‚˜ ์œ„ํ˜‘์— ๊ตดํ•˜์ง€ ์•Š์„ ๊ฒƒโ€์ด๋ผ๋Š” ์†Œ์‹ ์„ ์ง€์ผฐ๋‹ค.

์ •์น˜์  ๋‚ด๋Ÿฌํ‹ฐ๋ธŒ๊ฐ€ ๋งŒ๋“ค์–ด๋‚ธ ์„ธ๊ธˆ ์‹ ํ™”

RLE๊ฐ€ ์ง€๋‚œ 12์›” ๋งจํ•ดํŠผ ์—ฐ๊ตฌ์†Œ(Manhattan Institute)์˜ ์‹œํ‹ฐ ์ €๋„(City Journal)์— ๋ฐœํ‘œํ•œ โ€œ์ƒ์œ„ 10๊ฐ€์ง€ ์„ธ๊ธˆ ์‹ ํ™” ์ˆ˜์ •ํ•˜๊ธฐ(Correcting the Top 10 Tax Myths)โ€œ๋ผ๋Š” ๊ธ€์€ ํญ๋„“์€ ๋ฐ˜ํ–ฅ์„ ์ผ์œผ์ผฐ๋‹ค. ๊ทธ๋…€๋Š” ์ด ๊ธ€์„ ์“ฐ๊ฒŒ ๋œ ๋™๊ธฐ์— ๋Œ€ํ•ด โ€œ์‚ฌ๋žŒ๋“ค์ด ์ž˜๋ชป๋œ ์ •๋ณด๋ฅผ ๊ฐ€์ง€๊ณ  ๋…ผ์Ÿํ•˜๋Š” ๊ฒƒ์„ ๋ณด๊ณ  ์ •๋ง ํ™”๊ฐ€ ๋‚ฌ๋‹คโ€๊ณ  ๋งํ•œ๋‹ค. ํŠนํžˆ 2025๋…„์ด ์„ธ๊ธˆ ์ •์ฑ…์— ์žˆ์–ด ๊ฐ€์žฅ ์ค‘์š”ํ•œ ํ•ด๊ฐ€ ๋  ๊ฒƒ์ด๋ผ๋Š” ์˜ˆ์ธก ์†์—์„œ, ๋” ํ˜„๋ช…ํ•œ ๊ตญ๊ฐ€์  ๋…ผ์˜๋ฅผ ์œ„ํ•œ ๋ฐฐ๊ฒฝ ์ง€์‹์„ ์ œ๊ณตํ•˜๊ณ ์ž ํ–ˆ๋‹ค.

RLE๋Š” ์„ธ๊ธˆ์— ๋Œ€ํ•œ โ€œ์ž˜๋ชป๋œ ๋‚ด๋Ÿฌํ‹ฐ๋ธŒ(false narratives)โ€œ๊ฐ€ ๋ณด์ˆ˜์™€ ์ง„๋ณด ์–‘์ธก ๋ชจ๋‘์—์„œ ๋‚˜์˜จ๋‹ค๊ณ  ์ง€์ ํ•œ๋‹ค.

  • ๋ณด์ˆ˜์ฃผ์˜ ์ง„์˜์˜ ์ž˜๋ชป๋œ ์ฃผ์žฅ: ์„ธ๊ธˆ ๊ฐ๋ฉด์˜ ๊ธ์ •์ ์ธ ์ธก๋ฉด์„ ๊ณผ๋Œ€ํ‰๊ฐ€ํ•œ๋‹ค. โ€œ์„ธ๊ธˆ ๊ฐ๋ฉด์€ ์Šค์Šค๋กœ ์žฌ์›์„ ๋งˆ๋ จํ•œ๋‹ค(tax cuts pay for themselves)โ€œ๊ฑฐ๋‚˜ โ€œ์„ธ๊ธˆ ๊ฐ๋ฉด์ด ์ง€์ถœ์„ ์ค„์—ฌ ์ •๋ถ€๋ฅผ ๊ตถ์ฃผ๋ฆฌ๊ฒŒ ํ•  ๊ฒƒ(tax cuts will starve the beast)โ€œ์ด๋ผ๋Š” ์ฃผ์žฅ์ด ๋Œ€ํ‘œ์ ์ด๋‹ค. RLE๋Š” ์ด๋Ÿฌํ•œ ์ฃผ์žฅ์ด ์„ธ๊ธˆ ๊ฐ๋ฉด์˜ โ€œ๋งˆ๋ฒ• ๊ฐ™์€ ํž˜โ€์— ๋Œ€ํ•œ ๊ณผ๋„ํ•œ ๊ณผ์žฅ์ด๋ผ๊ณ  ๋น„ํŒํ•œ๋‹ค.
  • ์ง„๋ณด์ฃผ์˜ ์ง„์˜์˜ ์ž˜๋ชป๋œ ์ฃผ์žฅ: โ€œ์ค‘์‚ฐ์ธต์ด ๋ชจ๋“  ์„ธ๊ธˆ์„ ๋‚ด๊ณ , ๋Œ€๊ธฐ์—…๊ณผ ๋ถ€์œ ์ธต์€ ์•„๋ฌด๊ฒƒ๋„ ๋‚ด์ง€ ์•Š๋Š”๋‹คโ€๋Š” ์ฃผ์žฅ, โ€œ์ ์ž๋Š” ์„ธ๊ธˆ ๊ฐ๋ฉด ๋•Œ๋ฌธ์ด๋ฉฐ, ์œ ๋Ÿฝ์ฒ˜๋Ÿผ ๋ถ€์ž์™€ ๊ธฐ์—…์— ๋†’์€ ์„ธ๊ธˆ์„ ๋ถ€๊ณผํ•˜๋ฉด ํ•ด๊ฒฐ๋  ์ˆ˜ ์žˆ๋‹คโ€๋Š” ์ฃผ์žฅ์ด ์ฃผ๋ฅผ ์ด๋ฃฌ๋‹ค. RLE๋Š” ์ด๋Ÿฌํ•œ ์ฃผ์žฅ์ด โ€œ๋‹น์‹ ์€ ์ฐฉ์ทจ๋‹นํ•˜๊ณ  ์žˆ๋‹คโ€๋Š” ๋Œ€์ค‘์ด ๋“ฃ๊ณ  ์‹ถ์–ด ํ•˜๋Š” ์ด์•ผ๊ธฐ๋ฅผ ๋“ค๋ ค์ฃผ๋ฉฐ, โ€œํฐ์†๋“ค์„ ๊ณค๊ฒฝ์— ๋น ๋œจ๋ฆฌ๋ฉด ๋‹น์‹ ์—๊ฒŒ๋Š” ์•„๋ฌด๋Ÿฐ ์˜ํ–ฅ์„ ๋ฏธ์น˜์ง€ ์•Š๊ณ  ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•  ์ˆ˜ ์žˆ๋‹คโ€๋Š” ํŽธ๋ฆฌํ•œ ๋‚ด๋Ÿฌํ‹ฐ๋ธŒ๋ฅผ ์ œ๊ณตํ•œ๋‹ค๊ณ  ์„ค๋ช…ํ•œ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์‹ค์ œ ์ˆ˜์น˜๋Š” ์ด ๋‚ด๋Ÿฌํ‹ฐ๋ธŒ๊ฐ€ โ€œ์—„์ฒญ๋‚˜๊ฒŒ ๊ณผ์žฅ๋˜์—ˆ๋‹คโ€๋Š” ๊ฒƒ์„ ๋ช…ํ™•ํžˆ ๋ณด์—ฌ์ค€๋‹ค๊ณ  ๊ฐ•์กฐํ•œ๋‹ค.

์ด๋Ÿฌํ•œ ์˜คํ•ด๋Š” ๊ถ๊ทน์ ์œผ๋กœ ์ •์น˜์ธ๋“ค์ด ์œ ๊ถŒ์ž๋“ค์—๊ฒŒ โ€œ๊ณต์งœ ์ ์‹ฌ(free lunch)โ€œ์„ ์•ฝ์†ํ•˜๊ธฐ ๋•Œ๋ฌธ์— ๋ฐœ์ƒํ•œ๋‹ค. ๊ณตํ™”๋‹น์€ ๋‚ฎ์€ ์„ธ๊ธˆ์œผ๋กœ ๋งŽ์€ ์„ธ์ˆ˜์™€ ์ง€์ถœ ์‚ญ๊ฐ์„ ์•ฝ์†ํ•˜๊ณ , ์ง„๋ณด์ฃผ์˜์ž๋“ค์€ ๋‹ค๋ฅธ ๋ชจ๋“  ์‚ฌ๋žŒ๋“ค์—๊ฒŒ ์„ธ๊ธˆ์„ ๋ถ€๊ณผํ•˜์—ฌ ๋งŽ์€ ์ง€์ถœ์„ ํ•  ์ˆ˜ ์žˆ๋‹ค๊ณ  ์•ฝ์†ํ•œ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ํ˜„์‹ค์€ ๊ณตํ™”๋‹น์ด ์ง‘๊ถŒํ•˜๋ฉด ๋ฏผ์ฃผ๋‹น ๋ชป์ง€์•Š๊ฒŒ, ๋•Œ๋กœ๋Š” ๋” ๋งŽ์ด ์ง€์ถœํ•˜๊ณ , ๋ฏผ์ฃผ๋‹น์€ ์„ธ๊ธˆ ์ธ์ƒ์„ ์•ฝ์†ํ•˜์ง€๋งŒ ์‹ค์ œ๋กœ ํฌ๊ฒŒ ์ธ์ƒํ•˜์ง€ ๋ชปํ•ด ์ง€์ถœ๋งŒ ๋Š˜๋ฆฌ๋Š” ๊ฒฝํ–ฅ์ด ์žˆ๋‹ค๋Š” ๊ฒƒ์ด RLE์˜ ์ง€์ ์ด๋‹ค.

๋ฏธ๊ตญ ์„ธ๊ธˆ ์‹œ์Šคํ…œ์˜ 10๊ฐ€์ง€ ์‹ ํ™” ํŒŒํ—ค์น˜๊ธฐ

์ œ์‹œ์นด RLE๋Š” ๋ฏธ๊ตญ ์„ธ๊ธˆ ์‹œ์Šคํ…œ์— ๋Œ€ํ•œ ๊ฐ€์žฅ ํ”ํ•œ 10๊ฐ€์ง€ ์˜คํ•ด๋ฅผ ๋‹ค์Œ๊ณผ ๊ฐ™์ด ๋ฐ”๋กœ์žก๋Š”๋‹ค.

  1. ์„ธ๊ธˆ ๊ฐ๋ฉด์€ ์Šค์Šค๋กœ ์žฌ์›์„ ๋งˆ๋ จํ•œ๋‹ค: ์„ธ๊ธˆ ๊ฐ๋ฉด์€ ์ถ”๊ฐ€ ์„ธ์ˆ˜๋ฅผ ๊ฐ€์ ธ์˜ฌ ์ˆ˜ ์žˆ์ง€๋งŒ, ๊ฑฐ์˜ ํ•ญ์ƒ ์Šค์Šค๋กœ ์žฌ์›์„ ๋งˆ๋ จํ•˜์ง€ ๋ชปํ•œ๋‹ค.
  2. ์„ธ๊ธˆ ๊ฐ๋ฉด์€ ์ง€์ถœ์„ ์ค„์—ฌ ์ •๋ถ€๋ฅผ ๊ตถ์ฃผ๋ฆฌ๊ฒŒ ํ•œ๋‹ค: ์—ญ์‚ฌ์ ์œผ๋กœ๋Š” ๊ทธ ๋ฐ˜๋Œ€๋‹ค. ์„ธ๊ธˆ์„ ๊ฐ๋ฉดํ•˜๋ฉด ์˜ํšŒ๋Š” ์ง€์ถœ์„ ๋Š˜๋ฆฌ๊ณ , ์„ธ๊ธˆ์„ ์ธ์ƒํ•˜๋ฉด ์ง€์ถœ์„ ์ค„์ด๋Š” ๊ฒฝํ–ฅ์ด ์žˆ๋‹ค.
  3. ์ค‘์‚ฐ์ธต์ด ๋ถ€์œ ์ธต๋ณด๋‹ค ๋†’์€ ์„ธ์œจ์„ ๋ถ€๋‹ดํ•œ๋‹ค: ์‚ฌ์‹ค์ด ์•„๋‹ˆ๋‹ค. ๋ชจ๋“  ์—ฐ๋ฐฉ์„ธ๋ฅผ ํ•ฉ์‚ฐํ•˜๋ฉด ์ƒ์œ„ 1%๋Š” 33%, ์ค‘์‚ฐ์ธต์€ 12%, ํ•˜์œ„์ธต์€ ๊ฑฐ์˜ 0%๋ฅผ ๋ถ€๋‹ดํ•œ๋‹ค. ๋ฏธ๊ตญ์€ OECD(๊ฒฝ์ œํ˜‘๋ ฅ๊ฐœ๋ฐœ๊ธฐ๊ตฌ) ๊ตญ๊ฐ€ ์ค‘ ๊ฐ€์žฅ ๋ˆ„์ง„์ ์ธ(progressive) ์กฐ์„ธ ์‹œ์Šคํ…œ์„ ๊ฐ€์ง€๊ณ  ์žˆ๋‹ค.
  4. 1950๋…„๋Œ€์˜ 91% ์„ธ์œจ์ด ๋ง‰๋Œ€ํ•œ ์„ธ์ˆ˜๋ฅผ ์ฐฝ์ถœํ–ˆ๋‹ค: ๋‹น์‹œ 91% ์„ธ์œจ์„ ์‹ค์ œ๋กœ ๋‚ธ ์‚ฌ๋žŒ์€ ๊ฑฐ์˜ ์—†์—ˆ๋‹ค. 50%๋ฅผ ๋„˜๋Š” ์„ธ์œจ์„ ๋‚ธ ์‚ฌ๋žŒ๋„ ๊ฑฐ์˜ ์—†์—ˆ์œผ๋ฉฐ, ์ด๋“ค ์ตœ๊ณ  ์„ธ์œจ ๊ตฌ๊ฐ„์ด ๊ฑฐ๋‘ฌ๋“ค์ธ ์„ธ์ˆ˜๋Š” ์‚ฌ์‹ค์ƒ ๋ฏธ๋ฏธํ–ˆ๋‹ค.
  5. ์œ ๋Ÿฝ์€ ๋ถ€์ž๋“ค์—๊ฒŒ ๋” ๋งŽ์€ ์„ธ๊ธˆ์„ ๋ถ€๊ณผํ•˜์—ฌ ๋” ํฐ ์ •๋ถ€๋ฅผ ์šด์˜ํ•œ๋‹ค: ์‹ค์ œ๋กœ๋Š” ๋ถ€์œ ์ธต์— ๋Œ€ํ•œ ์„ธ๊ธˆ์€ ๋ฏธ๊ตญ๊ณผ ๋น„์Šทํ•˜๋‹ค. ์œ ๋Ÿฝ์˜ ์ถ”๊ฐ€ ์„ธ์ˆ˜๋Š” ์ฃผ๋กœ ์ค‘์‚ฐ์ธต์—๊ฒŒ ๋ถ€๋‹ด์„ ์ฃผ๋Š” ๋ถ€๊ฐ€๊ฐ€์น˜์„ธ(VAT, Value Added Tax)์™€ ๊ฐ™์€ ์ „๊ตญ์ ์ธ ํŒ๋งค์„ธ์—์„œ ๋น„๋กฏ๋œ๋‹ค.
  6. ๋ถ€์ž๋“ค์„ ์œ„ํ•œ ์„ธ๊ธˆ ๊ฐ๋ฉด์ด ๋Œ€๊ทœ๋ชจ ์žฌ์ • ์ ์ž์˜ ์›์ธ์ด๋‹ค: 2000๋…„ ์ดํ›„ ์„ธ๊ธˆ์€ GDP์˜ 2%p ๊ฐ์†Œํ–ˆ๋Š”๋ฐ, ์ด ์ค‘ ๋ถ€์œ ์ธต์— ๋Œ€ํ•œ ๊ฐ์„ธ๋Š” 0.6%p์— ๋ถˆ๊ณผํ•˜๋‹ค. ๋ฐ˜๋ฉด ์ง€์ถœ์€ GDP์˜ 6%p ์ฆ๊ฐ€ํ–ˆ๋‹ค. ์ง€์ถœ ์ฆ๊ฐ€๊ฐ€ ํ›จ์”ฌ ํฐ ์›์ธ์ด๋‹ค.
  7. ๊ธฐ์—…๊ณผ ๋ฐฑ๋งŒ์žฅ์ž์—๊ฒŒ ์„ธ๊ธˆ์„ ๋ถ€๊ณผํ•˜๋ฉด ์ ์ž๋ฅผ ์—†์•จ ์ˆ˜ ์žˆ๋‹ค: ์ด๋“ค์—๊ฒŒ 100% ์„ธ๊ธˆ์„ ๋ถ€๊ณผํ•˜๊ณ  ๋ชจ๋“  ์žฌ์‚ฐ์„ ๋ชฐ์ˆ˜ํ•˜๋”๋ผ๋„ ์ ์ž๋ฅผ ํ•ด๊ฒฐํ•˜๋Š” ๋ฐ๋Š” ํ„ฑ์—†์ด ๋ถ€์กฑํ•˜๋‹ค.
  8. 2017๋…„ ์„ธ๊ธˆ ๊ฐ๋ฉด์˜ ๋Œ€๋ถ€๋ถ„์€ ๊ธฐ์—…๊ณผ ๋ถ€์œ ์ธต์—๊ฒŒ ๋Œ์•„๊ฐ”๋‹ค: ์ˆœ์ˆ˜ํ•œ ๊ธˆ์•ก ๋ฉด์—์„œ๋Š” ๋” ๋งŽ์€ ๊ฐ์„ธ๋ฅผ ๋ฐ›์•˜์ง€๋งŒ, ๊ทธ๋“ค์ด ๋‚ด๋˜ ์„ธ๊ธˆ์˜ ๋น„์œจ๋กœ ๋ณด๋ฉด ๋Œ€๋žต ๋น„๋ก€์ ์ธ ์†Œ๋“์„ธ ๊ฐ๋ฉด์ด์—ˆ๋‹ค. ๋ชจ๋‘๊ฐ€ ์•ฝ 1%p ์ •๋„์˜ ์„ธ์œจ ์ธํ•˜ ํ˜œํƒ์„ ๋ฐ›์•˜๋‹ค.
  9. 1980๋…„ ์„ธ๋ฒ•์œผ๋กœ ๋Œ์•„๊ฐ€๋ฉด(๋ ˆ์ด๊ฑด, ๋ถ€์‹œ, ํŠธ๋Ÿผํ”„ ๊ฐ์„ธ ์ฒ ํšŒ) ๊ณ ํ†ต ์—†๋Š” ์ ์ž ๊ฐ์†Œ๊ฐ€ ๊ฐ€๋Šฅํ•˜๋‹ค: ๊ทธ๋ ‡๊ฒŒ ํ•˜๋ฉด ์ค‘์‚ฐ์ธต์˜ ์„ธ๊ธˆ ๋ถ€๋‹ด์ด ์—„์ฒญ๋‚˜๊ฒŒ ์ฆ๊ฐ€ํ•˜์—ฌ ๋ฐ›์•„๋“ค์ผ ์ˆ˜ ์—†๋Š” ์ˆ˜์ค€์ด ๋  ๊ฒƒ์ด๋‹ค.
  10. ๋ฏธ๊ตญ์˜ ๋ฒ•์ธ์„ธ๋Š” ๊ตญ์ œ ๊ธฐ์ค€์— ํ›จ์”ฌ ๋ชป ๋ฏธ์นœ๋‹ค: ๋ฏธ๊ตญ์€ 2017๋…„๊นŒ์ง€ ์„ ์ง„๊ตญ ์ค‘ ๊ฐ€์žฅ ๋†’์€ ๋ฒ•์ธ์„ธ์œจ์„ ๊ฐ€์ง€๊ณ  ์žˆ์—ˆ๋‹ค. 2017๋…„ ๋ฒ•์ธ์„ธ ์ธํ•˜ ์ดํ›„์—๋„ ๋ฏธ๊ตญ์˜ ๋ฒ•์ • ๋ฐ ์‹คํšจ ๋ฒ•์ธ์„ธ์œจ์€ ์—ฌ์ „ํžˆ ์ƒ์œ„ 1/3์— ์†ํ•œ๋‹ค.

์›Œ๋ Œ ๋ฒ„ํ•(Warren Buffett)์ด ์ž์‹ ์˜ ๋น„์„œ๋ณด๋‹ค ๋‚ฎ์€ ์„ธ์œจ์„ ๋‚ธ๋‹ค๊ณ  ๋งํ•œ ์œ ๋ช…ํ•œ ์ผํ™” ์—ญ์‹œ ์˜คํ•ด์˜ ์†Œ์ง€๊ฐ€ ์žˆ๋‹ค. ๊ทธ์˜ ์ˆ˜์ž… ๋Œ€๋ถ€๋ถ„์ด ํˆฌ์ž ์ˆ˜์ต์ธ ์ž๋ณธ ์ด๋“(capital gains) ํ˜•ํƒœ์ด๊ธฐ ๋•Œ๋ฌธ์ธ๋ฐ, ์ด๋Š” ๋งค๊ฐํ•  ๋•Œ๊นŒ์ง€ ๊ณผ์„ธ๋˜์ง€ ์•Š๋Š”๋‹ค. ํ•˜์ง€๋งŒ RLE๋Š” ์‹ค์ œ ๋ฐ์ดํ„ฐ๋ฅผ ๋ณด๋ฉด ์ž๋ณธ ์ด๋“์„ธ๋ฅผ ๊ณ ๋ คํ•˜๋”๋ผ๋„ ๊ณ ์†Œ๋“์ž๊ฐ€ ์ €์†Œ๋“์ž๋ณด๋‹ค ํ›จ์”ฌ ๋†’์€ ์„ธ์œจ์„ ๋‚ธ๋‹ค๊ณ  ์„ค๋ช…ํ•œ๋‹ค. ๋ฐ”์ด๋“  ํ–‰์ •๋ถ€๊ฐ€ ๋ถ€์œ ์ธต์˜ ์„ธ๊ธˆ ๋ถ€๋‹ด์„ 8%๋กœ ๊ณ„์‚ฐํ•œ ๋ฐฉ์‹์— ๋Œ€ํ•ด์„œ๋„, RLE๋Š” โ€œ์ƒ๋‹นํžˆ ๋ถ€์ •์งํ–ˆ๋‹คโ€๊ณ  ๋น„ํŒํ•œ๋‹ค. ๋ฏธ์‹คํ˜„๋œ ๋ถ€(unrealized wealth)๋ฅผ ์†Œ๋“์œผ๋กœ ๊ฐ„์ฃผํ•˜๊ณ , ๋ฒ•์ธ์„ธ ๋ฐ ์ƒ์†์„ธ ๋“ฑ ๋ถ€์œ ์ธต์ด ๋ถ€๋‹ดํ•˜๋Š” ๋‹ค๋ฅธ ์„ธ๊ธˆ์„ ๊ณ„์‚ฐ์— ํฌํ•จํ•˜์ง€ ์•Š์•˜๋‹ค๋Š” ๊ฒƒ์ด๋‹ค.

โ€˜๊ณ ๊ท€ํ•œ ๊ฑฐ์ง“๋งโ€™์˜ ์œ ํ˜น๊ณผ ์žฌ์ •์  ๋ฌด์ฑ…์ž„

RLE๋Š” ์›Œ์‹ฑํ„ด์˜ ๊ฒฝ์ œ ์ •์ฑ… ์ž…์•ˆ์ž๋“ค์ด โ€œ์ง€์  ์ •์ง์„ฑโ€์„ ๊ฐ€์ง€๊ณ  ๊ฒฝ๋ ฅ์„ ์Œ“๋‹ค๊ฐ€๋„ ๊ถŒ๋ ฅ์˜ ์œ ํ˜น ์•ž์—์„œ โ€œ์†๋ฌผ(hacks)โ€œ์ด ๋˜๋Š” ๊ฒƒ์„ ๋ชฉ๊ฒฉํ–ˆ๋‹ค๊ณ  ๋งํ•œ๋‹ค. โ€œ์ƒ๋Œ€๋ฐฉ๋„ ๋”๋Ÿฝ๊ฒŒ ๋†€ ํ…Œ๋‹ˆ ์šฐ๋ฆฌ๋„ ๊ทธ๋ž˜์•ผ ํ•œ๋‹คโ€๋Š” ์ธ์‹๊ณผ โ€œ๊ฒฐ๊ตญ ์˜ฌ๋ฐ”๋ฅธ ์ •์ฑ…์œผ๋กœ ์ด๋Œ โ€˜๊ณ ๊ท€ํ•œ ๊ฑฐ์ง“๋ง(noble lie)โ€˜โ€œ์ด๋ผ๋Š” ์ž๊ธฐํ•ฉ๋ฆฌํ™”๊ฐ€ ๋งŒ์—ฐํ•ด ์žˆ๋‹ค๋Š” ๊ฒƒ์ด๋‹ค.

์ด๋Ÿฌํ•œ ์˜คํ•ด์™€ ์กฐ์ž‘์˜ ๊ถ๊ทน์ ์ธ ๋Œ€๊ฐ€๋Š” โ€œ์žฌ์•™์ ์ธ ์„ธ๋ฒ•โ€์œผ๋กœ ์ด์–ด์ง„๋‹ค. ํ˜„์žฌ ๋ฏธ๊ตญ์˜ ์„ธ๋ฒ•์€ ๊ทน๋„๋กœ ๋ณต์žกํ•˜๊ณ  ๋น„ํšจ์œจ์ ์ด๋ฉฐ, ์ง€์ถœ์„ ๊ฐ๋‹นํ•  ๋งŒํผ ์ถฉ๋ถ„ํ•œ ์„ธ์ˆ˜๋ฅผ ๊ฑฐ๋‘์ง€ ๋ชปํ•˜๊ณ  ์žˆ๋‹ค. RLE์— ๋”ฐ๋ฅด๋ฉด, 2000๋…„ ์ดํ›„ ์ ์ž ์ฆ๊ฐ€์˜ ์•ฝ 3๋ถ„์˜ 1์€ ์„ธ๊ธˆ ์ •์ฑ…, 3๋ถ„์˜ 2๋Š” ์ง€์ถœ ์ •์ฑ… ๋•Œ๋ฌธ์ด๋ผ๊ณ  ๋ถ„์„ํ•œ๋‹ค.

ํ˜„์žฌ ๋ฏธ๊ตญ์˜ ๊ตญ๊ฐ€ ๋ถ€์ฑ„๋Š” ์•ฝ 39์กฐ ๋‹ฌ๋Ÿฌ(์•ฝ 5๊ฒฝ 3์ฒœ์กฐ ์›)์— ๋‹ฌํ•˜๋ฉฐ, GDP ๋Œ€๋น„ 124%๋กœ ์ œ2์ฐจ ์„ธ๊ณ„๋Œ€์ „ ์งํ›„ ์ดํ›„ ์ตœ๊ณ  ์ˆ˜์ค€์ด๋‹ค. ์ด ๋ถ€์ฑ„์— ๋Œ€ํ•œ ์ด์ž๋งŒ ํ•ด๋„ 2021๋…„ 3,500์–ต ๋‹ฌ๋Ÿฌ์—์„œ ํ˜„์žฌ ๊ฑฐ์˜ 1์กฐ ๋‹ฌ๋Ÿฌ๋กœ 3๋ฐฐ ๊ฐ€๊นŒ์ด ์ฆ๊ฐ€ํ–ˆ์œผ๋ฉฐ, 10๋…„ ํ›„์—๋Š” 2์กฐ ๋‹ฌ๋Ÿฌ์— ์ด๋ฅผ ๊ฒƒ์œผ๋กœ ์˜ˆ์ƒ๋œ๋‹ค. ์ด์ž๋Š” ์ด๋ฏธ ๋ฉ”๋””์ผ€์ด๋“œ(Medicaid), ๊ตญ๋ฐฉ, ๋ฉ”๋””์ผ€์–ด(Medicare)๋ฅผ ๋„˜์–ด ์‚ฌํšŒ๋ณด์žฅ(Social Security) ๋‹ค์Œ์œผ๋กœ ํฐ ์˜ˆ์‚ฐ ํ•ญ๋ชฉ์ด ๋˜์—ˆ๊ณ , 2042๋…„์—๋Š” ์‚ฌํšŒ๋ณด์žฅ๋งˆ์ € ์ถ”์›”ํ•  ๊ฒƒ์œผ๋กœ ๋ณด์ธ๋‹ค. RLE๋Š” ์ด ๋ชจ๋“  ๋ฌธ์ œ์˜ ์ฃผ๋ฒ”์œผ๋กœ โ€œ๋ˆ ์“ฐ๋Š” ๊ฒƒ์„ ์ ˆ๋Œ€ ๋ฉˆ์ถœ ์ˆ˜ ์—†๋Š” ์—ฐ๋ฐฉ ์ •๋ถ€โ€๋ฅผ ์ง€๋ชฉํ•œ๋‹ค. 2000๋…„ ์ดํ›„ ์–‘๋‹น ๋ชจ๋‘ ๋Œ€๊ทœ๋ชจ ๊ฐ์„ธ์™€ ์ง€์ถœ ํ™•๋Œ€๋ฅผ ์•ฝ์†ํ•˜๋Š” โ€œ๊ตฐ๋น„ ๊ฒฝ์Ÿ(arms race)โ€œ์— ๋Œ์ž…ํ–ˆ์œผ๋ฉฐ, ์ด๋Š” ์ง€๋‚œ 50๋…„ ๋™์•ˆ ์ „๋ก€ ์—†๋Š” ์ง€์ถœ ์ฆ๊ฐ€๋กœ ์ด์–ด์กŒ๋‹ค.

ํ•ด๊ฒฐ์ฑ…: ์‚ฌํšŒ๋ณด์žฅ์ œ๋„ ๊ฐœํ˜๊ณผ ์ค‘์‚ฐ์ธต ์ฆ์„ธ๋ผ๋Š” โ€˜์„ธ ๋ฒˆ์งธ ๋ ˆ์ผโ€™

RLE๊ฐ€ ์ œ์‹œํ•˜๋Š” ํ•ด๊ฒฐ์ฑ…์€ ๋‘ ๊ฐ€์ง€, ์ฆ‰ โ€œ๋งค์šฐ ๊ณจ์น˜ ์•„ํ”ˆ(particularly vexing)โ€ ๋ฌธ์ œ์™€ ์ง๊ฒฐ๋œ๋‹ค. ๋ฐ”๋กœ ์‚ฌํšŒ๋ณด์žฅ์ œ๋„ ๋ฐ ๋ฉ”๋””์ผ€์–ด์™€ ๊ฐ™์€ ์‚ฌํšŒ๋ณด์žฅ์ œ๋„ ๊ฐœํ˜๊ณผ ์ค‘์‚ฐ์ธต ์ฆ์„ธ์ด๋‹ค.

์‚ฌํšŒ๋ณด์žฅ์ œ๋„์™€ ๋ฉ”๋””์ผ€์–ด๋Š” ์„ธ๊ธˆ์œผ๋กœ ์Šค์Šค๋กœ ์žฌ์›์„ ์กฐ๋‹ฌํ•˜์ง€ ๋ชปํ•œ๋‹ค๋Š” ๊ฒƒ์ด RLE์˜ ์ฃผ์žฅ์ด๋‹ค. ์ด๋“ค์€ ๋ง‰๋Œ€ํ•œ ์ ์ž๋ฅผ ๋‚ด๊ณ  ์žˆ์œผ๋ฉฐ, ํ–ฅํ›„ 30๋…„๊ฐ„ 124์กฐ ๋‹ฌ๋Ÿฌ์˜ ํ˜„๊ธˆ ๋ถ€์กฑ์•ก์ด ๋ฐœ์ƒํ•  ๊ฒƒ์œผ๋กœ ์˜ˆ์ƒ๋œ๋‹ค. ํ˜„์žฌ์˜ ์„ธ๊ธˆ์€ ํ˜„์žฌ ์ˆ˜ํ˜œ์ž๋“ค์—๊ฒŒ ์ง€๊ธ‰๋  ๋ฟ, ๋ฏธ๋ž˜๋ฅผ ์œ„ํ•ด ์ €์ถ•๋˜์ง€ ์•Š๋Š”๋‹ค. ์ˆ˜ํ˜œ์ž๊ฐ€ ๋Š˜์–ด๋‚˜๋ฉด ์„ธ๊ธˆ์„ ๋‚ด๋Š” ์‚ฌ๋žŒ๋งŒ์œผ๋กœ๋Š” ๊ฐ๋‹นํ•  ์ˆ˜ ์—†๊ฒŒ ๋œ๋‹ค. ์ „ํ˜•์ ์ธ ๋…ธ์ธ๋“ค์€ ๋ฉ”๋””์ผ€์–ด์—์„œ ์ž์‹ ์ด ๋‚ฉ๋ถ€ํ•œ ๊ธˆ์•ก์˜ 3๋ฐฐ๋ฅผ ๋Œ๋ ค๋ฐ›๊ณ  ์žˆ๋‹ค๋Š” ์ถฉ๊ฒฉ์ ์ธ ๋ฐ์ดํ„ฐ๋„ ์ œ์‹œ๋œ๋‹ค.

์ด๋Ÿฌํ•œ ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•œ ์‚ฌํšŒ๋ณด์žฅ์ œ๋„ ๊ฐœํ˜์—๋Š” ์„ธ ๊ฐ€์ง€ ์ง€๋ ›๋Œ€(levers)๊ฐ€ ์žˆ๋‹ค. ๋ฐ”๋กœ ์„ธ๊ธˆ ์ธ์ƒ, ์ˆ˜๊ธ‰ ์—ฐ๋ น ์ƒํ–ฅ, ๊ณ ์†Œ๋“์ธต์— ๋Œ€ํ•œ ํ˜œํƒ ๊ฐœํ˜์ด๋‹ค. ํŠนํžˆ RLE๋Š” ๊ณ ์†Œ๋“์ธต์— ๋Œ€ํ•œ ํ˜œํƒ ๊ฐœํ˜, ์ฆ‰ **์†Œ๋“ ์‹ฌ์‚ฌ(means testing)**๋ฅผ ํ†ตํ•ด ์—ฐ๊ฐ„ 6๋งŒ ๋‹ฌ๋Ÿฌ ์ด์ƒ์˜ ์‚ฌํšŒ๋ณด์žฅ ํ˜œํƒ์„ ๋ฐ›๋Š” ๋ฐฑ๋งŒ์žฅ์ž๋“ค์˜ ํ˜œํƒ์„ ์ค„์—ฌ์•ผ ํ•œ๋‹ค๊ณ  ์ฃผ์žฅํ•œ๋‹ค. ์‚ฌํšŒ๋ณด์žฅ์ œ๋„๋Š” ๋นˆ๊ณค ์˜ˆ๋ฐฉ ํ”„๋กœ๊ทธ๋žจ์ด์ง€, ๋ชจ๋‘๋ฅผ ๋ถ€์œ ํ•˜๊ฒŒ ํ•˜๋Š” ๋ณดํŽธ์ ์ธ ํ˜œํƒ์ด ์•„๋‹ˆ๋ผ๋Š” ๊ฒƒ์ด๋‹ค.

๊ทธ๋Ÿฌ๋‚˜ ์ด๋Ÿฌํ•œ ๊ฐœํ˜์€ ์›Œ์‹ฑํ„ด์—์„œ โ€œ์„ธ ๋ฒˆ์งธ ๋ ˆ์ผ(third rail, ๋งŒ์ง€๋ฉด ๊ฐ์ „๋˜๋Š” ์œ„ํ—˜ํ•œ ๊ฒƒ)โ€œ๋กœ ์—ฌ๊ฒจ์ง„๋‹ค. ๋Œ€์ค‘์€ ์‚ฌํšŒ๋ณด์žฅ์„ธ๊ฐ€ 100% ์ ๋ฆฝ๋˜์–ด ์ž์‹ ์˜ ๊ณ„์ขŒ์— ์Œ“์—ฌ ์žˆ๋‹ค๊ณ  ๋ฏฟ๊ธฐ ๋•Œ๋ฌธ์—, ์–ด๋–ค ์‚ญ๊ฐ๋„ โ€œ๋ถ„๋…ธ์™€ ๋„๋‘‘์งˆโ€๋กœ ๊ฐ„์ฃผํ•œ๋‹ค. RLE๋Š” ์ •์น˜์ธ๋“ค์ด ๊ฐœ์ธ์ ์œผ๋กœ๋Š” ์ด๋Ÿฌํ•œ ๊ฐœํ˜์˜ ํ•„์š”์„ฑ์— ๋™์˜ํ•˜์ง€๋งŒ, โ€œ์ •์น˜์ ์œผ๋กœ๋Š” ๋ถˆ๊ฐ€๋Šฅํ•˜๋‹คโ€๋ฉฐ ๊ณต๊ฐœ์ ์œผ๋กœ ๋‚˜์„œ๋Š” ๊ฒƒ์„ ์ฃผ์ €ํ•œ๋‹ค๊ณ  ์ง€์ ํ•œ๋‹ค. ๋ฐ”์ด๋“ ์ด๋‚˜ ํŠธ๋Ÿผํ”„ ๊ฐ™์€ ๋Œ€ํ†ต๋ น์ด ์นด๋ฉ”๋ผ ์•ž์—์„œ ์†”์งํ•˜๊ฒŒ ์ ์ž ๋ฌธ์ œ์™€ ์‚ฌํšŒ๋ณด์žฅ ํ˜œํƒ ์‚ญ๊ฐ์˜ ํ•„์š”์„ฑ์„ ์—ญ์„คํ•˜๋”๋ผ๋„, ๋‹นํŒŒ์  ์ดํ•ด๊ด€๊ณ„์™€ ๋Œ€์ค‘์˜ ์˜คํ•ด ๋•Œ๋ฌธ์— ์ง€์ง€๋ฐ›๊ธฐ ์–ด๋ ค์šธ ๊ฒƒ์ด๋ผ๋Š” ๋น„๊ด€์ ์ธ ์ „๋ง์„ ๋‚ด๋†“๋Š”๋‹ค.

๊ฒฐ๊ตญ RLE๋Š” ์ด๋Ÿฌํ•œ ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๋ ค๋ฉด โ€œ์–‘๋‹น์ด ์‚ฌ์ ์œผ๋กœ ๋ชจ์—ฌ ๋ชจ๋“  ๊ฒƒ์„ ๋…ผ์˜ํ•ด์•ผ ํ•œ๋‹คโ€๊ณ  ์ฃผ์žฅํ•œ๋‹ค. ์„ธ๊ธˆ๊ณผ ์ง€์ถœ, ์‚ฌํšŒ๋ณด์žฅ์ œ๋„ ๊ฐœํ˜์„ ๋ชจ๋‘ ํ•จ๊ป˜ ๋‹ค๋ฃจ์–ด ๋ชจ๋‘๊ฐ€ ํ•จ๊ป˜ ๋…ธ๋ ฅํ•˜๊ณ  ํฌ์ƒํ•ด์•ผ ํ•œ๋‹ค๋Š” ๊ฒƒ์ด๋‹ค. ํ•˜์ง€๋งŒ ๊ทธ๋…€๋Š” ํ˜„์žฌ์˜ ์ •์น˜ ํ™˜๊ฒฝ์—์„œ๋Š” ์ด๋Ÿฌํ•œ ์ผ์ด ์ผ์–ด๋‚˜๊ธฐ๋ฅผ โ€œ์ˆจ์ฃฝ์—ฌ ๊ธฐ๋‹ค๋ฆด ์ˆ˜ ์—†๋‹คโ€๊ณ  ๋ง๋ถ™์ด๋ฉฐ, ๋ฏธ๊ตญ์˜ ์žฌ์ • ๋ฏธ๋ž˜์— ๋Œ€ํ•œ ๊นŠ์€ ์šฐ๋ ค๋ฅผ ํ‘œ๋ช…ํ•œ๋‹ค.


โ€œAI Agents & Claude Skills Full Course: Setup and Buildโ€ โ€” Greg Isenberg ๊ธฐ๋ฐ˜ ๊ธฐ์‚ฌ ์›๋ณธ ์˜์ƒ ๋ณด๊ธฐ

AI ์—์ด์ „ํŠธ ํ™œ์šฉ์˜ ๋น„๋ฐ€: โ€˜์ปจํ…์ŠคํŠธโ€™์™€ โ€˜์Šคํ‚ฌโ€™๋กœ ์ƒ์‚ฐ์„ฑ์„ ๊ทน๋Œ€ํ™”ํ•˜๋Š” ๋ฐฉ๋ฒ•

ํ˜„์žฌ ์ธ๊ณต์ง€๋Šฅ(AI) ๊ธฐ์ˆ ์€ ๋ˆˆ๋ถ€์‹  ๋ฐœ์ „์„ ๊ฑฐ๋“ญํ•˜๋ฉฐ ์šฐ๋ฆฌ์˜ ์—…๋ฌด ํ™˜๊ฒฝ์„ ํ˜์‹ ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ํด๋กœ๋“œ ์˜คํ‘ธ์Šค(Claude Opus) 4.6์ด๋‚˜ GPT 5.4์™€ ๊ฐ™์€ ๋Œ€๊ทœ๋ชจ ์–ธ์–ด ๋ชจ๋ธ(LLM)๋“ค์€ ์ธ๊ฐ„์˜ ์–ธ์–ด๋ฅผ ๋†€๋ž๋„๋ก ์ •ํ™•ํ•˜๊ฒŒ ์ดํ•ดํ•˜๊ณ  ์ƒ์„ฑํ•˜๋ฉฐ, ์ด์ œ๋Š” ๋ณต์žกํ•œ ์ž‘์—…์„ ์ž๋™ํ™”ํ•˜๋Š” AI ์—์ด์ „ํŠธ(AI Agent)์˜ ์‹œ๋Œ€๊ฐ€ ๋„๋ž˜ํ–ˆ์Šต๋‹ˆ๋‹ค. ํ•˜์ง€๋งŒ ๋งŽ์€ ์‚ฌ์šฉ์ž๋“ค์ด ๊ฐ•๋ ฅํ•œ AI ๋ชจ๋ธ์„ ์ œ๋Œ€๋กœ ํ™œ์šฉํ•˜์ง€ ๋ชปํ•˜๊ณ , ์˜คํžˆ๋ ค ๋น„ํšจ์œจ๊ณผ ์ขŒ์ ˆ์„ ๊ฒฝํ—˜ํ•˜๊ณ  ์žˆ๋‹ค๊ณ  AI ์—์ด์ „ํŠธ ์ „๋ฌธ๊ฐ€ ๋กœ์Šค ๋งˆ์ดํฌ(Ross Mike)๋Š” ์ง€์ ํ•ฉ๋‹ˆ๋‹ค. ๊ทธ์˜ ํ•ต์‹ฌ ์ฃผ์žฅ์€ ๊ฐ„๋‹จํ•ฉ๋‹ˆ๋‹ค: ๋ชจ๋ธ์€ ์ด๋ฏธ ๋›ฐ์–ด๋‚˜์ง€๋งŒ, โ€˜์ปจํ…์ŠคํŠธ(Context)โ€˜๊ฐ€ ๋ชจ๋“  ๊ฒƒ์„ ๊ฒฐ์ •ํ•˜๋ฉฐ, ์šฐ๋ฆฌ๋Š” ์—์ด์ „ํŠธ์™€ ์ƒํ˜ธ์ž‘์šฉํ•˜๋Š” ๋ฐฉ์‹์„ ๊ทผ๋ณธ์ ์œผ๋กœ ์žฌ๊ณ ํ•ด์•ผ ํ•œ๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค.

1. ๋ชจ๋ธ์€ ์ด๋ฏธ ๋›ฐ์–ด๋‚˜๋‹ค, ํ•ต์‹ฌ์€ โ€˜์ปจํ…์ŠคํŠธโ€™

๊ณผ๊ฑฐ์—๋Š” ๋ชจ๋ธ ์„ฑ๋Šฅ ์ž์ฒด์— ๋Œ€ํ•œ ๋…ผ์Ÿ์ด ์žˆ์—ˆ์ง€๋งŒ, ์ด์ œ๋Š” ํด๋กœ๋“œ ์˜คํ‘ธ์Šค๋‚˜ GPT ๊ฐ™์€ ์ตœ์‹  ๋ชจ๋ธ๋“ค์ด โ€˜์˜ˆ์™ธ์ ์œผ๋กœ ํ›Œ๋ฅญํ•˜๋‹คโ€™๋Š” ๊ฒƒ์ด ๋กœ์Šค ๋งˆ์ดํฌ์˜ ์ฃผ์žฅ์ž…๋‹ˆ๋‹ค. ํŠนํžˆ ํ”„๋กœ๊ทธ๋ž˜๋ฐ ๋ถ„์•ผ์—์„œ๋Š” ํŠน์ • ๋ชจ๋ธ์ด UI ๋””์ž์ธ์—, ๋‹ค๋ฅธ ๋ชจ๋ธ์ด ๋ฐฑ์—”๋“œ ๊ฐœ๋ฐœ์— ๋” ์ ํ•ฉํ•˜๋‹ค๋Š” ์ธ์‹์ด ์žˆ์ง€๋งŒ, ์ „๋ฐ˜์ ์œผ๋กœ ๋ชจ๋ธ๋“ค์€ ์ด๋ฏธ ๋งค์šฐ ๋†’์€ ์ˆ˜์ค€์— ๋„๋‹ฌํ–ˆ์Šต๋‹ˆ๋‹ค.

ํ•˜์ง€๋งŒ ์—ฌ๊ธฐ์„œ ์ค‘์š”ํ•œ ๊ฒƒ์€ **์ปจํ…์ŠคํŠธ(Context)**์ž…๋‹ˆ๋‹ค. ๋ชจ๋ธ์€ ๋‹น์‹ ์ด ์ œ๊ณตํ•˜๋Š” ์ปจํ…์ŠคํŠธ์— ๋”ฐ๋ผ โ€˜์ตœ๊ณ ์˜ ๊ฒฐ๊ณผ๋ฌผโ€™์„ ๋‚ด๋†“์„ ์ˆ˜๋„ ์žˆ๊ณ , โ€˜์—‰๋ง์ง„์ฐฝ(slop)โ€˜์˜ ๊ฒฐ๊ณผ๋ฌผ์„ ๋‚ด๋†“์„ ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค. ์ปจํ…์ŠคํŠธ๋ž€ ๋ชจ๋ธ์ด ํŠน์ • ์ž‘์—…์„ ์‹คํ–‰ํ•˜๋Š” ๋ฐ ํ•„์š”ํ•œ ์ •๋ณด๋ฅผ ์ทจํ•ฉํ•˜๋Š” ๋ฐฉ์‹์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด, ์ฝ”๋”ฉ ์—์ด์ „ํŠธ์˜ ์ปจํ…์ŠคํŠธ ์ฐฝ์€ ๋‹ค์Œ๊ณผ ๊ฐ™์€ ์š”์†Œ๋“ค๋กœ ๊ตฌ์„ฑ๋ฉ๋‹ˆ๋‹ค:

  • ์‹œ์Šคํ…œ ํ”„๋กฌํ”„ํŠธ(System Prompt): ๋ชจ๋ธ ์ œ๊ณต์ž๊ฐ€ ์„ค์ •ํ•œ ์ผ๋ฐ˜์ ์ธ ์ง€์นจ์œผ๋กœ, ๋ชจ๋ธ์ด ์–ด๋–ป๊ฒŒ ํ–‰๋™ํ•˜๊ณ  ๋ฌด์—‡์„ ํ•ด์•ผ ํ•˜๋Š”์ง€ ์•ˆ๋‚ดํ•ฉ๋‹ˆ๋‹ค. (์˜ˆ: ์ตœ๊ทผ ์œ ์ถœ๋œ ํด๋กœ๋“œ ์ฝ”๋“œ์˜ ์‹œ์Šคํ…œ ํ”„๋กฌํ”„ํŠธ๋Š” ๊ฐœ๋ฐœ์ž๋“ค์—๊ฒŒ ๋ชจ๋ธ์˜ ์ž‘๋™ ๋ฐฉ์‹์„ ์ดํ•ดํ•˜๋Š” ๋ฐ ํฐ ํ†ต์ฐฐ๋ ฅ์„ ์ฃผ์—ˆ์Šต๋‹ˆ๋‹ค.)
  • ์—์ด์ „ํŠธ.md ๋˜๋Š” ํด๋กœ๋“œ.md ํŒŒ์ผ (Agent.md/Claude.md files): ์‚ฌ์šฉ์ž๊ฐ€ ์—์ด์ „ํŠธ์—๊ฒŒ ํŠน์ • ์ •๋ณด๋ฅผ ์ œ๊ณตํ•˜๋Š” ์ •์  ํŒŒ์ผ์ž…๋‹ˆ๋‹ค.
  • ์Šคํ‚ฌ(Skills): ์—์ด์ „ํŠธ๊ฐ€ ํ•„์š”ํ•  ๋•Œ๋งˆ๋‹ค ๋™์ ์œผ๋กœ ์ฐธ์กฐํ•˜๋Š” ์ •๋ณด ๋ฌถ์Œ์ž…๋‹ˆ๋‹ค.
  • ๋„๊ตฌ(Tools): ์ฝ๊ธฐ, ์“ฐ๊ธฐ ๋“ฑ ์—์ด์ „ํŠธ๊ฐ€ ์™ธ๋ถ€์™€ ์ƒํ˜ธ์ž‘์šฉํ•˜๋Š” ๋ฐ ์‚ฌ์šฉํ•˜๋Š” ๋‚ด์žฅ ๋„๊ตฌ๋“ค์ž…๋‹ˆ๋‹ค.
  • ์ฝ”๋“œ๋ฒ ์ด์Šค(Codebase): ์—์ด์ „ํŠธ๊ฐ€ ์ž‘์—…ํ•˜๋Š” ์ฝ”๋“œ ์ž์ฒด์ž…๋‹ˆ๋‹ค.
  • ์‚ฌ์šฉ์ž ๋Œ€ํ™”(User Conversation): ์‚ฌ์šฉ์ž์™€ ์—์ด์ „ํŠธ ๊ฐ„์˜ ๋ชจ๋“  ์ƒํ˜ธ์ž‘์šฉ ๊ธฐ๋ก์ž…๋‹ˆ๋‹ค.

์ด ๋ชจ๋“  ์š”์†Œ๋“ค์ด ์ปจํ…์ŠคํŠธ ์ฐฝ์„ ์ฑ„์šฐ๋ฉฐ, ๋Œ€ํ™”๊ฐ€ ๊ธธ์–ด์งˆ์ˆ˜๋ก ์ปจํ…์ŠคํŠธ ์ฐฝ์€ ํ™•์žฅ๋˜์–ด ์ตœ๋Œ€ ํ† ํฐ(token) ํ•œ๊ณ„(์˜ˆ: 250,000 ํ† ํฐ)์— ๋„๋‹ฌํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ปจํ…์ŠคํŠธ ์ฐฝ์ด ๋„ˆ๋ฌด ๋งŽ์€ ์ •๋ณด๋กœ ์ฑ„์›Œ์ง€๋ฉด ๋ชจ๋ธ์˜ ์„ฑ๋Šฅ์ด ์ €ํ•˜๋˜๊ณ , ๋ถˆํ•„์š”ํ•œ ํ† ํฐ ์‚ฌ์šฉ์œผ๋กœ ๋น„์šฉ๋„ ์ฆ๊ฐ€ํ•ฉ๋‹ˆ๋‹ค.

2. ์—์ด์ „ํŠธ.md ํŒŒ์ผ์€ ์ด์ œ ๊ทธ๋งŒ! โ€˜์Šคํ‚ฌโ€™์˜ ์ง„์ •ํ•œ ํž˜

๋งŽ์€ ์‚ฌ๋žŒ๋“ค์ด ์—์ด์ „ํŠธ์˜ ํ–‰๋™ ๋ฐฉ์นจ์„ ์ •์˜ํ•˜๊ธฐ ์œ„ํ•ด ์—์ด์ „ํŠธ.md๋‚˜ ํด๋กœ๋“œ.md(claude.md) ํŒŒ์ผ์„ ์‚ฌ์šฉํ•˜์ง€๋งŒ, ๋กœ์Šค ๋งˆ์ดํฌ๋Š” 95%์˜ ๊ฒฝ์šฐ ์ด๋Ÿฌํ•œ ํŒŒ์ผ์ด โ€˜๋ถˆํ•„์š”ํ•˜๋‹คโ€™๊ณ  ๋‹จ์–ธํ•ฉ๋‹ˆ๋‹ค. ๊ทธ ์ด์œ ๋Š” ๋ชจ๋ธ์ด ์ด๋ฏธ ๋งค์šฐ ๋›ฐ์–ด๋‚˜๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด, ํด๋กœ๋“œ ์ฝ”๋“œ์—๊ฒŒ โ€œ์ด ์ฝ”๋“œ๋ฒ ์ด์Šค๋Š” React๋ฅผ ์‚ฌ์šฉํ•œ๋‹คโ€๊ณ  ์•Œ๋ ค์ค„ ํ•„์š”๊ฐ€ ์—†์Šต๋‹ˆ๋‹ค. ๋ชจ๋ธ์€ ์ฝ”๋“œ๋ฒ ์ด์Šค ์ž์ฒด๋ฅผ ์ปจํ…์ŠคํŠธ๋กœ ๊ฐ€์ง€๊ณ  ์žˆ์œผ๋ฏ€๋กœ, ์ฝ”๋“œ๋ฅผ ์ง์ ‘ ํ™•์ธํ•˜์—ฌ ์–ด๋–ค ํ”„๋ ˆ์ž„์›Œํฌ๊ฐ€ ์‚ฌ์šฉ๋˜์—ˆ๋Š”์ง€ ํŒŒ์•…ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

์—์ด์ „ํŠธ.md ํŒŒ์ผ์˜ ๊ทผ๋ณธ์ ์ธ ๋ฌธ์ œ๋Š” ๋ชจ๋“  ๋Œ€ํ™” ํ„ด๋งˆ๋‹ค ์ „์ฒด ๋‚ด์šฉ์ด ์ปจํ…์ŠคํŠธ์— ์ถ”๊ฐ€๋œ๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๋งŒ์•ฝ 1,000์ค„์งœ๋ฆฌ ์—์ด์ „ํŠธ.md ํŒŒ์ผ์ด 7,000ํ† ํฐ์„ ์ฐจ์ง€ํ•œ๋‹ค๋ฉด, ๋งค๋ฒˆ ์—์ด์ „ํŠธ์™€ ๋Œ€ํ™”ํ•  ๋•Œ๋งˆ๋‹ค 7,000ํ† ํฐ์„ ๋‚ญ๋น„ํ•˜๋Š” ์…ˆ์ž…๋‹ˆ๋‹ค. ์ด๋Š” ๋น„์šฉ ๋‚ญ๋น„์ผ ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ, ์ปจํ…์ŠคํŠธ ์ฐฝ์„ ๋น ๋ฅด๊ฒŒ ์ฑ„์›Œ ๋ชจ๋ธ์˜ ์„ฑ๋Šฅ์„ ์ €ํ•˜์‹œํ‚ต๋‹ˆ๋‹ค.

๋ฐ˜๋ฉด, ์Šคํ‚ฌ(Skills)์€ โ€˜์ ์ง„์  ๊ณต๊ฐœ(progressive disclosure)โ€™ ๋ฐฉ์‹์œผ๋กœ ์ž‘๋™ํ•ฉ๋‹ˆ๋‹ค. ์Šคํ‚ฌ ํŒŒ์ผ์—๋Š” ์ด๋ฆ„(name)๊ณผ ์„ค๋ช…(description), ๊ทธ๋ฆฌ๊ณ  ๊ทธ ์•„๋ž˜์— ์ƒ์„ธํ•œ ์ •๋ณด๊ฐ€ ๋‹ด๊ฒจ ์žˆ์Šต๋‹ˆ๋‹ค. ์ปจํ…์ŠคํŠธ์—๋Š” ์ฒ˜์Œ์—๋Š” ์Šคํ‚ฌ์˜ ์ด๋ฆ„๊ณผ ์„ค๋ช…๋งŒ ์ถ”๊ฐ€๋ฉ๋‹ˆ๋‹ค. ์—์ด์ „ํŠธ๊ฐ€ ํŠน์ • ์Šคํ‚ฌ์ด ํ•„์š”ํ•˜๋‹ค๊ณ  ํŒ๋‹จํ•  ๋•Œ๋งŒ ์ „์ฒด ์Šคํ‚ฌ ๋ฌธ์„œ์˜ ๋‚ด์šฉ์ด ์ปจํ…์ŠคํŠธ์— ์ถ”๊ฐ€๋ฉ๋‹ˆ๋‹ค.

์˜ˆ๋ฅผ ๋“ค์–ด, โ€˜๋…ธ์…˜ ๋ณด๊ณ ์„œ ์Šคํ‚ฌ(Notion Report Skill)โ€˜์ด ์žˆ๋‹ค๋ฉด, ์—์ด์ „ํŠธ์—๊ฒŒ โ€œ๋…ธ์…˜ ๋ณด๊ณ ์„œ๋ฅผ ๋งŒ๋“ค์–ด์ค˜โ€๋ผ๊ณ  ์ง€์‹œํ–ˆ์„ ๋•Œ, ์—์ด์ „ํŠธ๋Š” ์ปจํ…์ŠคํŠธ์— ์žˆ๋Š” ์Šคํ‚ฌ ๋ชฉ๋ก์—์„œ ์ด๋ฆ„๊ณผ ์„ค๋ช…์„ ๋ณด๊ณ  ์ด ์Šคํ‚ฌ์ด ํ•„์š”ํ•˜๋‹ค๊ณ  ํŒ๋‹จํ•ฉ๋‹ˆ๋‹ค. ๊ทธ์ œ์„œ์•ผ ๋…ธ์…˜ ๋ณด๊ณ ์„œ ์Šคํ‚ฌ์˜ ์ƒ์„ธ ๋‚ด์šฉ์ด ์ปจํ…์ŠคํŠธ์— ์ถ”๊ฐ€๋˜๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค. ์ด๋Š” ์ˆ˜์ฒœ ํ† ํฐ์„ ๋งค๋ฒˆ ์ถ”๊ฐ€ํ•˜๋Š” ๋Œ€์‹ , ์ˆ˜์‹ญ ํ† ํฐ๋งŒ์œผ๋กœ ํšจ์œจ์ ์ธ ์ž‘๋™์„ ๊ฐ€๋Šฅํ•˜๊ฒŒ ํ•ฉ๋‹ˆ๋‹ค.

๊ทธ๋ ‡๋‹ค๋ฉด ์—์ด์ „ํŠธ.md ํŒŒ์ผ์€ ์–ธ์ œ ์‚ฌ์šฉํ•ด์•ผ ํ• ๊นŒ์š”? ๋กœ์Šค ๋งˆ์ดํฌ๋Š” ํšŒ์‚ฌ ๊ณ ์œ ์˜ ๊ธฐ๋ฐ€ ์ •๋ณด๋‚˜ ํŠน์ • ๋ฐฉ๋ฒ•๋ก ์ฒ˜๋Ÿผ, ๋ชจ๋“  ๋Œ€ํ™”์—์„œ ๋ฐ˜๋“œ์‹œ ์ฐธ์กฐ๋˜์–ด์•ผ ํ•˜๋Š” ๋…์ ์ ์ธ ์ •๋ณด๊ฐ€ ์žˆ์„ ๋•Œ๋งŒ ์‚ฌ์šฉํ•ด์•ผ ํ•œ๋‹ค๊ณ  ์กฐ์–ธํ•ฉ๋‹ˆ๋‹ค. ํ•˜์ง€๋งŒ 95%์˜ ์‚ฌ๋žŒ๋“ค์€ ์ด๋Ÿฌํ•œ ํ•„์š”์„ฑ์„ ๋А๋ผ์ง€ ๋ชปํ•ฉ๋‹ˆ๋‹ค.

3. โ€˜์Šคํ‚ฌโ€™ ๊ตฌ์ถ•์˜ ํ™ฉ๊ธˆ๋ฅ : ์—์ด์ „ํŠธ์™€ ํ•จ๊ป˜ ๋ฐฐ์šฐ๊ณ  ๋ฐ˜๋ณตํ•˜๋ผ

๋งŽ์€ ์ด๋“ค์ด ์›Œํฌํ”Œ๋กœ๋ฅผ ํŒŒ์•…ํ•˜์ž๋งˆ์ž ๊ณง๋ฐ”๋กœ ์Šคํ‚ฌ์„ ๋งŒ๋“ค๋ ค ํ•˜์ง€๋งŒ, ๋กœ์Šค ๋งˆ์ดํฌ๋Š” ์ด๋ฅผ **โ€˜์ตœ์•…์˜ ์ ‘๊ทผ ๋ฐฉ์‹โ€™**์ด๋ผ๊ณ  ๊ฒฝ๊ณ ํ•ฉ๋‹ˆ๋‹ค. ์ด๋Š” ๋งˆ์น˜ ์‹ ์ž…์‚ฌ์›์„ ์ฑ„์šฉํ–ˆ์„ ๋•Œ ๋ฐ”๋กœ ๋ชจ๋“  ์—…๋ฌด๋ฅผ ๋งก๊ธฐ๊ณ  โ€˜์•Œ์•„์„œ ํ•ด๋ผโ€™๋ผ๊ณ  ํ•˜๋Š” ๊ฒƒ๊ณผ ๊ฐ™์Šต๋‹ˆ๋‹ค. ์‚ฌ๋žŒ์€ ์‹คํŒจ๋ฅผ ํ†ตํ•ด ๋ฐฐ์šฐ๊ณ , ๋ฉ˜ํ† ์˜ ์ง€๋„๋ฅผ ํ†ตํ•ด ์„ฑ์žฅํ•ฉ๋‹ˆ๋‹ค. AI ์—์ด์ „ํŠธ๋„ ๋งˆ์ฐฌ๊ฐ€์ง€์ž…๋‹ˆ๋‹ค.

๋กœ์Šค ๋งˆ์ดํฌ๊ฐ€ ์ œ์‹œํ•˜๋Š” ์Šคํ‚ฌ ๊ตฌ์ถ•์˜ ์˜ฌ๋ฐ”๋ฅธ ๋ฐฉ๋ฒ•์€ ๋‹ค์Œ๊ณผ ๊ฐ™์Šต๋‹ˆ๋‹ค:

  1. ์›Œํฌํ”Œ๋กœ ์‹๋ณ„ (Identify Workflow): ํŠน์ • ์ž‘์—…์„ ์ž๋™ํ™”ํ•ด์•ผ ํ•  ํ•„์š”์„ฑ์„ ๋А๋‚๋‹ˆ๋‹ค. (์˜ˆ: ์œ ํŠœ๋ธŒ ์Šคํฐ์„œ ์ด๋ฉ”์ผ ์„ ๋ณ„ ์ž‘์—…)
  2. ์—์ด์ „ํŠธ์™€ ๋‹จ๊ณ„๋ณ„ ์ƒํ˜ธ์ž‘์šฉ (Teach Step-by-Step): ์›Œํฌํ”Œ๋กœ๋ฅผ ์ฒ˜์Œ๋ถ€ํ„ฐ ๋๊นŒ์ง€ ์—์ด์ „ํŠธ์™€ ํ•จ๊ป˜ ์ˆ˜ํ–‰ํ•ฉ๋‹ˆ๋‹ค.
    • ๋กœ์Šค ๋งˆ์ดํฌ์˜ ์Šคํฐ์„œ ์ด๋ฉ”์ผ ์„ ๋ณ„ ์˜ˆ์‹œ:
      • ์ฒ˜์Œ์—๋Š” ์—์ด์ „ํŠธ์—๊ฒŒ โ€œ์ด๋ฉ”์ผ์„ ํ™•์ธํ•˜๊ณ  ์Šคํฐ์„œ์— ๋Œ€ํ•ด ์กฐ์‚ฌํ•ด์„œ ๊ฐ€์น˜๊ฐ€ ์žˆ๋Š”์ง€ ์•Œ๋ ค์ค˜โ€๋ผ๊ณ ๋งŒ ์ง€์‹œํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ฒฐ๊ณผ๋Š” ๋ชจ๋“  ์Šคํฐ์„œ๋ฅผ โ€˜ํ›Œ๋ฅญํ•˜๋‹คโ€™๊ณ  ํ‰๊ฐ€ํ•˜๋Š” ๊ฒƒ์ด์—ˆ์Šต๋‹ˆ๋‹ค.
      • ๋ชจ๋ธ์€ โ€˜์ƒ๊ฐโ€™ํ•˜๋Š” ๊ฒƒ์ด ์•„๋‹ˆ๋ผ ํ† ํฐ์„ ์˜ˆ์ธกํ•  ๋ฟ์ด๋ฏ€๋กœ, ๋ช…ํ™•ํ•œ ์ง€์นจ์ด ์—†์œผ๋ฉด ํ”ผ์ƒ์ ์ธ ๋‹ต๋ณ€์„ ๋‚ด๋†“์Šต๋‹ˆ๋‹ค.
      • ๋กœ์Šค ๋งˆ์ดํฌ๋Š” ์—์ด์ „ํŠธ์—๊ฒŒ โ€œ์ด๋ฉ”์ผ์„ ๋ณด๋ƒˆ์œผ๋‹ˆ, ํšŒ์‚ฌ ํŠธ์œ„ํ„ฐ, ์œ ํŠœ๋ธŒ, ํŠธ๋Ÿฌ์ŠคํŠธํŒŒ์ผ๋Ÿฟ(Trust Pilot)์„ ํ™•์ธํ•˜๊ณ  ํˆฌ์ž ์œ ์น˜ ์—ฌ๋ถ€๋ฅผ ํ™•์ธํ•ด. ์ด ์ค‘ ๋‘ ๊ฐ€์ง€ ์ด์ƒ์ด ์—†๊ฑฐ๋‚˜ ์ƒํƒœ๊ฐ€ ์ข‹์ง€ ์•Š์œผ๋ฉด ์ž๋™ ๊ฑฐ์ ˆ์ด์•ผโ€๋ผ๊ณ  ๋‹จ๊ณ„๋ณ„๋กœ ์ง€์‹œํ–ˆ์Šต๋‹ˆ๋‹ค.
      • ์—์ด์ „ํŠธ๋Š” ์ด ์ง€์นจ์— ๋”ฐ๋ผ ์„ฑ๊ณต์ ์œผ๋กœ ์Šคํฐ์„œ๋ฅผ ๊ฑฐ์ ˆํ•˜๊ณ  ๊ตฌ๊ธ€ ์‹œํŠธ์— ๊ธฐ๋กํ–ˆ์Šต๋‹ˆ๋‹ค.
  3. ์„ฑ๊ณต์ ์ธ ์‹คํ–‰ ๊ฒฝํ—˜ ์ถ•์  (Accumulate Successful Runs): ์—์ด์ „ํŠธ๊ฐ€ ์›Œํฌํ”Œ๋กœ๋ฅผ ์„ฑ๊ณต์ ์œผ๋กœ ์ˆ˜ํ–‰ํ•  ๋•Œ๊นŒ์ง€ ์—ฌ๋Ÿฌ ๋ฒˆ ๋ฐ˜๋ณตํ•˜๊ณ  ์ˆ˜์ •ํ•ฉ๋‹ˆ๋‹ค.
  4. AI์—๊ฒŒ ์Šคํ‚ฌ ์ƒ์„ฑ ์ง€์‹œ (Instruct AI to Create Skill): ์—์ด์ „ํŠธ๊ฐ€ ์›Œํฌํ”Œ๋กœ๋ฅผ ์™„๋ฒฝํ•˜๊ฒŒ ์ˆ˜ํ–‰ํ•œ ํ›„์—, โ€œ๋ฐฉ๊ธˆ ์ˆ˜ํ–‰ํ•œ ์ž‘์—…์„ ๊ฒ€ํ† ํ•˜๊ณ  ์ด๋ฅผ ์Šคํ‚ฌ๋กœ ๋งŒ๋“ค์–ด๋ผโ€๊ณ  ์ง€์‹œํ•ฉ๋‹ˆ๋‹ค.

์ด๋Ÿฌํ•œ ๊ณผ์ •์„ ํ†ตํ•ด ์ƒ์„ฑ๋œ ์Šคํ‚ฌ์€ ์‹ค์ œ ์„ฑ๊ณต์ ์ธ ์‹คํ–‰์˜ ์ปจํ…์ŠคํŠธ๋ฅผ ๋‹ด๊ณ  ์žˆ๊ธฐ ๋•Œ๋ฌธ์— ํ›จ์”ฌ ๋” ํšจ๊ณผ์ ์ž…๋‹ˆ๋‹ค. ๋งŽ์€ ์‚ฌ๋žŒ๋“ค์ด AI์—๊ฒŒ ์Šคํ‚ฌ์„ ๋งŒ๋“ค๋ผ๊ณ  ์ง€์‹œํ•˜๋”๋ผ๋„, ์—์ด์ „ํŠธ๊ฐ€ ์„ฑ๊ณต์ ์ธ ์‹คํ–‰ ๊ฒฝํ—˜์ด ์—†๋‹ค๋ฉด ๋ถˆ์™„์ „ํ•˜๊ฑฐ๋‚˜ ๋น„ํšจ์œจ์ ์ธ ์Šคํ‚ฌ์„ ์ƒ์„ฑํ•  ๊ฐ€๋Šฅ์„ฑ์ด ๋†’์Šต๋‹ˆ๋‹ค.

๋˜ํ•œ, ๋กœ์Šค ๋งˆ์ดํฌ๋Š” ๋‹ค๋ฅธ ์‚ฌ๋žŒ์ด ๋งŒ๋“  ์Šคํ‚ฌ์„ ๋‹ค์šด๋กœ๋“œํ•˜๋Š” ๊ฒƒ์— ๋Œ€ํ•ด ๊ฐ•๋ ฅํ•˜๊ฒŒ ๊ฒฝ๊ณ ํ•ฉ๋‹ˆ๋‹ค. ์ฒซ์งธ, ๋ณด์•ˆ์ƒ์˜ ์œ„ํ—˜์ด ์žˆ์Šต๋‹ˆ๋‹ค. ๋ฌด์ž‘์œ„ ์Šคํ‚ฌ์€ ์ž ์žฌ์ ์ธ ๊ณต๊ฒฉ ๋ฒกํ„ฐ๊ฐ€ ๋  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๋‘˜์งธ, ๋‹น์‹ ์˜ ์—์ด์ „ํŠธ์—๊ฒŒ๋Š” โ€˜์„ฑ๊ณต์ ์ธ ์‹คํ–‰์˜ ์ปจํ…์ŠคํŠธโ€™๊ฐ€ ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค. ๋‹ค๋ฅธ ์‚ฌ๋žŒ์˜ ์Šคํ‚ฌ์€ ๋‹น์‹ ์˜ ํŠน์ • ์›Œํฌํ”Œ๋กœ์™€ ๋งž์ง€ ์•Š์„ ์ˆ˜ ์žˆ์œผ๋ฉฐ, ์—์ด์ „ํŠธ๊ฐ€ ๊ทธ ์Šคํ‚ฌ์„ ์ œ๋Œ€๋กœ ํ™œ์šฉํ•˜์ง€ ๋ชปํ•  ๊ฐ€๋Šฅ์„ฑ์ด ํฝ๋‹ˆ๋‹ค.

4. ์ƒ์‚ฐ์„ฑ์„ ์œ„ํ•œ ํ™•์žฅ, โ€˜๋ฉ‹โ€™์ด ์•„๋‹Œ โ€˜์‹ค์šฉโ€™์„ ์œ„ํ•ด

๋งŽ์€ ์‚ฌ์šฉ์ž๋“ค์ด โ€˜๋ฉ‹์ ธ ๋ณด์ด๋Š”โ€™ ์„ค์ •์„ ์œ„ํ•ด ์ฒ˜์Œ๋ถ€ํ„ฐ ์ˆ˜๋งŽ์€ ์„œ๋ธŒ ์—์ด์ „ํŠธ(sub-agents)์™€ ์Šคํ‚ฌ์„ ๊ตฌ์ถ•ํ•˜์ง€๋งŒ, ์ด๋Š” ์ƒ์‚ฐ์„ฑ ํ–ฅ์ƒ์— ์˜คํžˆ๋ ค ๋ฐฉํ•ด๊ฐ€ ๋œ๋‹ค๋Š” ์ง€์ ์ž…๋‹ˆ๋‹ค. ๋งˆ์น˜ ํŒ€ ๊ด€๋ฆฌ ๊ฒฝํ—˜์ด ์ „ํ˜€ ์—†๋Š” ์‚ฌ๋žŒ์ด ๊ฐ‘์ž๊ธฐ 10๋ช…์˜ ์ง์›์„ ๊ฑฐ๋А๋ฆฐ ํšŒ์‚ฌ๋ฅผ ์‹œ์ž‘ํ•˜๋Š” ๊ฒƒ๊ณผ ๊ฐ™์Šต๋‹ˆ๋‹ค. ์ด๋Š” ๋น„ํšจ์œจ์ ์ด๊ณ  ํ˜ผ๋ž€์Šค๋Ÿฌ์šธ ๋ฟ์ž…๋‹ˆ๋‹ค.

๋กœ์Šค ๋งˆ์ดํฌ๋Š” **์ƒ์‚ฐ์„ฑ์„ ์œ„ํ•œ ํ™•์žฅ(scaling for productivity)**์„ ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค.

  1. ํ•˜๋‚˜์˜ ์—์ด์ „ํŠธ๋กœ ์‹œ์ž‘ํ•ฉ๋‹ˆ๋‹ค. ์ด ๋ฉ”์ธ ์—์ด์ „ํŠธ๊ฐ€ ๋ชจ๋“  ์ž‘์—…์„ ์ˆ˜ํ–‰ํ•˜๋„๋ก ํ•ฉ๋‹ˆ๋‹ค.
  2. ํ•„์š”ํ•œ ์Šคํ‚ฌ์„ ์ ์ง„์ ์œผ๋กœ ๊ตฌ์ถ•ํ•ฉ๋‹ˆ๋‹ค. ์œ„์—์„œ ์„ค๋ช…ํ•œ ๋ฐฉ์‹์œผ๋กœ ์—์ด์ „ํŠธ์™€ ํ•จ๊ป˜ ์›Œํฌํ”Œ๋กœ๋ฅผ ๋งŒ๋“ค๊ณ  ์Šคํ‚ฌ์„ ์™„์„ฑํ•ฉ๋‹ˆ๋‹ค.
  3. ์›Œํฌํ”Œ๋กœ๊ฐ€ ๋ช…ํ™•ํ•ด์ง€๊ณ  ์Šคํ‚ฌ์ด ๊ฒฌ๊ณ ํ•ด์ง€๋ฉด, ๊ทธ๋•Œ ์„œ๋ธŒ ์—์ด์ „ํŠธ๋ฅผ ์ถ”๊ฐ€ํ•ฉ๋‹ˆ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด, ์Šคํฐ์„œ ๊ด€๋ฆฌ ์›Œํฌํ”Œ๋กœ๊ฐ€ ํ™•๋ฆฝ๋˜๋ฉด, ๋งˆ์ผ€ํŒ… ๊ด€๋ จ ๋ชจ๋“  ์ž‘์—…์„ ๋‹ด๋‹นํ•˜๋Š” ์„œ๋ธŒ ์—์ด์ „ํŠธ๋ฅผ ๋งŒ๋“ค ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

์ด๋Ÿฌํ•œ ๋ฐฉ์‹์€ ์ฒ˜์Œ์—๋Š” โ€˜๋œ ์„น์‹œํ•ดโ€™ ๋ณด์ผ ์ˆ˜ ์žˆ์ง€๋งŒ, ๊ฒฐ๊ตญ์—๋Š” ํ›จ์”ฌ ๋” ์ƒ์‚ฐ์ ์ด๊ณ  ํšจ์œจ์ ์ธ ์‹œ์Šคํ…œ์„ ๊ตฌ์ถ•ํ•  ์ˆ˜ ์žˆ๊ฒŒ ํ•ฉ๋‹ˆ๋‹ค. ์—์ด์ „ํŠธ๋ฅผ ๋งˆ์น˜ โ€˜์‹ ์ž… ์ง์›โ€™์ฒ˜๋Ÿผ ๋Œ€ํ•˜๊ณ , ๊ทธ๋“ค์„ ๊ฐ€๋ฅด์น˜๊ณ  ์„ฑ์žฅ์‹œํ‚ค๋Š” ๊ณผ์ •์—์„œ ์šฐ๋ฆฌ ์ž์‹ ๋„ AI๋ฅผ ํšจ๊ณผ์ ์œผ๋กœ ํ™œ์šฉํ•˜๋Š” โ€˜๊ฐœ์ธ์ ์ธ ์ธ๊ฐ„ ์Šคํ‚ฌโ€™์„ ๊ฐœ๋ฐœํ•˜๊ฒŒ ๋ฉ๋‹ˆ๋‹ค.

AI๊ฐ€ ํ™”์ดํŠธ์นผ๋ผ ์ง๋ฌด๋ฅผ ๋Œ€์ฒดํ•  ๊ฒƒ์ด๋ผ๋Š” โ€˜์˜๊ตฌ์ ์ธ ํ•˜์ธต ๊ณ„๊ธ‰(permanent underclass)โ€™ ๋‹ด๋ก ์ด ์žˆ์ง€๋งŒ, ๋กœ์Šค ๋งˆ์ดํฌ๋Š” ์ด๋Ÿฌํ•œ ๋„๊ตฌ์˜ ์ž‘๋™ ๋ฐฉ์‹๊ณผ ํ™œ์šฉ๋ฒ•์„ ์•„๋Š” ์‚ฌ๋žŒ๋“ค์—๊ฒŒ๋Š” ๋ฐ์€ ๋ฏธ๋ž˜๊ฐ€ ๊ธฐ๋‹ค๋ฆฌ๊ณ  ์žˆ๋‹ค๊ณ  ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Š” โ€œ20๋…„ ๊ฑธ๋ฆฌ๋˜ ์ง€์‹ ์Šต๋“์ด ์ด์ œ ์›” 20๋‹ฌ๋Ÿฌ๋ฉด ๊ฐ€๋Šฅํ•ด์กŒ๋‹คโ€๋ฉฐ, ๋น„๊ธฐ์ˆ ์ง ์‚ฌ๋žŒ๋“ค๋„ AI๋ฅผ ํ†ตํ•ด ์–ต๋Œ€ ์ˆ˜์ต์„ ์˜ฌ๋ฆฌ๋Š” ์‚ฌ๋ก€๋ฅผ ์–ธ๊ธ‰ํ•˜๋ฉฐ ์‹œ๋Œ€์˜ ๋ณ€ํ™”๋ฅผ ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค.

5. ์ฝ”๋“œ ํ”„๋กœ์ ํŠธ์™€ ํ…œํ”Œ๋ฆฟ์˜ ์žฌ์กฐ๋ช…

์ฝ”๋”ฉ ๋ถ„์•ผ์—์„œ๋„ ์—์ด์ „ํŠธ ํ™œ์šฉ ๋ฐฉ์‹์— ๋Œ€ํ•œ ํ†ต์ฐฐ์„ ์ œ์‹œํ•ฉ๋‹ˆ๋‹ค. ๋ชจ๋ธ๋“ค์€ ํŠนํžˆ ํƒ€์ž…์Šคํฌ๋ฆฝํŠธ(TypeScript)์™€ ๊ฐ™์€ ์ฝ”๋“œ๋ฅผ ์ž‘์„ฑํ•˜๋Š” ๋ฐ ํƒ์›”ํ•˜๋ฉฐ, ํด๋กœ๋“œ ์ฝ”์›Œํฌ(Claude Co-Work)๋‚˜ ์˜คํ”ˆ ํด๋กœ๋“œ(OpenClaude) ๊ฐ™์€ ๋„๊ตฌ๋“ค์€ ์ฝ”๋“œ๋ฅผ ์ž‘์„ฑํ•˜๊ณ  API๋ฅผ ํ˜ธ์ถœํ•˜๋Š” ๋ฐฉ์‹์œผ๋กœ ์ž‘๋™ํ•ฉ๋‹ˆ๋‹ค.

ํฅ๋ฏธ๋กœ์šด ์ ์€, ์ฝ”๋“œ ํ”„๋กœ์ ํŠธ๋ฅผ ๊ตฌ์ถ•ํ•  ๋•Œ๋„ ํŠน์ • ๊ธฐ์ˆ  ์Šคํƒ(tech stack)์— ๋Œ€ํ•œ ์ž์„ธํ•œ ์„ค๋ช…์„ ์—์ด์ „ํŠธ.md ํŒŒ์ผ์— ๋‹ด์„ ํ•„์š”๊ฐ€ ์—†๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค. ์ฝ”๋“œ๋ฒ ์ด์Šค ์ž์ฒด๊ฐ€ ์ปจํ…์ŠคํŠธ ์—ญํ• ์„ ํ•˜๋ฏ€๋กœ, ๋ชจ๋ธ์€ ์ฝ”๋“œ๋ฅผ ์ฝ๊ณ  ํ•„์š”ํ•œ ์ •๋ณด๋ฅผ ํŒŒ์•…ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

๋กœ์Šค ๋งˆ์ดํฌ๋Š” ๊ณผ๊ฑฐ ํ…œํ”Œ๋ฆฟ(template) ์‹œ์žฅ์ด ๋ฒˆ์„ฑํ–ˆ๋˜ ๊ฒƒ์ฒ˜๋Ÿผ, **โ€˜ํ…œํ”Œ๋ฆฟ์˜ ๋ฅด๋„ค์ƒ์Šค(renaissance)โ€˜**๊ฐ€ ์˜ฌ ๊ฒƒ์ด๋ผ๊ณ  ์ „๋งํ•ฉ๋‹ˆ๋‹ค. ์›น ์•ฑ์ด๋‚˜ ๋ชจ๋ฐ”์ผ ์•ฑ์„ ์œ„ํ•œ ๊ฒฌ๊ณ ํ•œ ๊ธฐ์ดˆ ํ…œํ”Œ๋ฆฟ์€ AI ์—์ด์ „ํŠธ์—๊ฒŒ ๊ฐ•๋ ฅํ•œ ์ดˆ๊ธฐ ์ปจํ…์ŠคํŠธ๋ฅผ ์ œ๊ณตํ•˜์—ฌ, ์—์ด์ „ํŠธ๊ฐ€ ๊ทธ ์œ„์— ํšจ์œจ์ ์œผ๋กœ ๊ธฐ๋Šฅ์„ ๊ตฌ์ถ•ํ•  ์ˆ˜ ์žˆ๊ฒŒ ํ•ฉ๋‹ˆ๋‹ค. ์ด ์—ญ์‹œ โ€˜์ตœ์†Œํ•œ์˜ ์ปจํ…์ŠคํŠธ ์‚ฌ์šฉโ€™์ด๋ผ๋Š” ์›์น™๊ณผ ์ผ๋งฅ์ƒํ†ตํ•ฉ๋‹ˆ๋‹ค.

6. ์Šคํ‚ฌ์˜ ์žฌ๊ท€์  ๊ตฌ์ถ•: ์‹คํŒจ๋ฅผ ํ†ตํ•œ ์™„๋ฒฝ

์Šคํ‚ฌ์„ ๋งŒ๋“ค์—ˆ๋‹ค๊ณ  ํ•ด์„œ ๋์ด ์•„๋‹™๋‹ˆ๋‹ค. ์—์ด์ „ํŠธ๊ฐ€ ์—ฌ์ „ํžˆ ์‹ค์ˆ˜๋ฅผ ์ €์ง€๋ฅผ ์ˆ˜ ์žˆ์œผ๋ฉฐ, ์ด๋•Œ๊ฐ€ ๋ฐ”๋กœ ์Šคํ‚ฌ์„ ๋”์šฑ ์™„๋ฒฝํ•˜๊ฒŒ ๋งŒ๋“ค ๊ธฐํšŒ์ž…๋‹ˆ๋‹ค. ๋กœ์Šค ๋งˆ์ดํฌ๋Š” ์ด๋ฅผ **โ€˜์Šคํ‚ฌ์˜ ์žฌ๊ท€์  ๊ตฌ์ถ•(recursively building skills)โ€˜**์ด๋ผ๊ณ  ๋ถ€๋ฆ…๋‹ˆ๋‹ค.

  1. ์‹คํŒจ ์‹๋ณ„: ์—์ด์ „ํŠธ๊ฐ€ ์ž‘์—…์„ ์‹คํŒจํ•˜๋ฉด ๋ถˆํ‰ํ•˜๋Š” ๋Œ€์‹ , โ€œ์™œ ์‹คํŒจํ–ˆ๋Š”์ง€?โ€, โ€œ์–ด๋–ค ์˜ค๋ฅ˜๊ฐ€ ๋ฐœ์ƒํ–ˆ๋Š”์ง€?โ€๋ฅผ ๋ฌผ์–ด๋ด…๋‹ˆ๋‹ค. ์—์ด์ „ํŠธ๋Š” ๋†€๋ž๋„๋ก ์ƒ์„ธํ•˜๊ฒŒ ์˜ค๋ฅ˜ ์›์ธ(์˜ˆ: 5005 ์˜ค๋ฅ˜, ํฌ๋ ˆ๋”ง ๋ถ€์กฑ)์„ ์•Œ๋ ค์ค„ ๊ฒƒ์ž…๋‹ˆ๋‹ค.
  2. ํ•ด๊ฒฐ์ฑ… ์ œ์‹œ: ์—์ด์ „ํŠธ์—๊ฒŒ ์˜ค๋ฅ˜๋ฅผ ์ˆ˜์ •ํ•˜๋„๋ก ์ง€์‹œํ•ฉ๋‹ˆ๋‹ค. (์˜ˆ: โ€œ์—ฌ๊ธฐ์— ๋ฌธ์ œ๊ฐ€ ์žˆ์—ˆ์œผ๋‹ˆ ๊ณ ์ณ๋ผ.โ€)
  3. ์Šคํ‚ฌ ์—…๋ฐ์ดํŠธ: ์—์ด์ „ํŠธ๊ฐ€ ์˜ค๋ฅ˜๋ฅผ ์„ฑ๊ณต์ ์œผ๋กœ ์ˆ˜์ •ํ•˜๋ฉด, โ€œ์ด ๋ฌธ์ œ๊ฐ€ ๋‹ค์‹œ ๋ฐœ์ƒํ•˜์ง€ ์•Š๋„๋ก ์Šคํ‚ฌ์„ ์—…๋ฐ์ดํŠธํ•˜๋ผโ€๊ณ  ์ง€์‹œํ•ฉ๋‹ˆ๋‹ค.

๋กœ์Šค ๋งˆ์ดํฌ๋Š” ์ด ๊ณผ์ •์„ 5๋ฒˆ ๋ฐ˜๋ณตํ•˜์—ฌ ์œ ํŠœ๋ธŒ ์ฑ„๋„์˜ ๋ณต์žกํ•œ ๋ณด๊ณ ์„œ ์ƒ์„ฑ ์Šคํ‚ฌ์„ ์™„์„ฑํ–ˆ๋‹ค๊ณ  ๋งํ•ฉ๋‹ˆ๋‹ค. ์ด ์Šคํ‚ฌ์€ ๋…ธ์…˜, ์œ ํŠœ๋ธŒ ๋ถ„์„, ํŠธ์œ„ํ„ฐ ๋ถ„์„ ๋“ฑ 8๊ฐœ์˜ ๋ฐ์ดํ„ฐ ์†Œ์Šค์—์„œ ์ •๋ณด๋ฅผ ๊ฐ€์ ธ์˜ค๋Š” ์ž‘์—…์ธ๋ฐ, ์ด์ œ๋Š” ๋‹จ ํ•œ ๋ฒˆ์˜ ํ”„๋กฌํ”„ํŠธ๋กœ 10๋ถ„ ๋งŒ์— ์™„๋ฒฝํ•˜๊ฒŒ ์‹คํ–‰๋ฉ๋‹ˆ๋‹ค. ์ด๋Ÿฌํ•œ ๋ฐ˜๋ณต์ ์ด๊ณ  ๊ฒฝํ—˜์ ์ธ ํ•™์Šต ๊ณผ์ •์„ ํ†ตํ•ด ์—์ด์ „ํŠธ๋Š” ์ ์  ๋” ๊ฐ•๋ ฅํ•˜๊ณ  ์ •ํ™•ํ•œ ์Šคํ‚ฌ์„ ๊ฐ–์ถ”๊ฒŒ ๋ฉ๋‹ˆ๋‹ค.

์ด ๊ณผ์ •์—๋Š” ์ดˆ๊ธฐ 2์ฃผ ์ •๋„์˜ โ€˜ํˆฌ์ž ์‹œ๊ฐ„โ€™์ด ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค. ์ฒ˜์Œ์—๋Š” ์—์ด์ „ํŠธ๊ฐ€ ํ˜ผ๋ž€์Šค๋Ÿฌ์›Œํ•˜๊ณ  โ€˜์“ฐ๋ ˆ๊ธฐโ€™ ๊ฐ™๋‹ค๊ณ  ๋А๋‚„ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ๋งˆ์น˜ ์‹ ์ž…์‚ฌ์›์—๊ฒŒ โ€œ์žฌ๋ฌด ๋ณด๊ณ ์„œ๋ฅผ ์ž‘์„ฑํ•ด๋ผโ€๊ณ  ํ–ˆ์„ ๋•Œ, ๊ทธ๊ฐ€ ์‚ฌ์—…์— ๋Œ€ํ•œ ์ปจํ…์ŠคํŠธ๊ฐ€ ์—†์–ด โ€œ์–ด๋””์„œ๋ถ€ํ„ฐ ์‹œ์ž‘ํ•ด์•ผ ํ• ์ง€ ๋ชจ๋ฅด๊ฒ ๋‹คโ€๊ณ  ํ•˜๋Š” ๊ฒƒ๊ณผ ๊ฐ™์Šต๋‹ˆ๋‹ค. ์ธ๊ธฐ ์‹œํŠธ์ฝค โ€˜์˜คํ”ผ์Šค(The Office)โ€˜์˜ ์ง(Jim)์ด โ€˜์š”์•ฝ ๋ณด๊ณ (rundown)โ€˜๊ฐ€ ๋ฌด์—‡์ธ์ง€ ๋ชฐ๋ผ ํ—ค๋งค๋˜ ์—ํ”ผ์†Œ๋“œ์ฒ˜๋Ÿผ, ์—์ด์ „ํŠธ๋„ ๋ช…ํ™•ํ•œ ์ปจํ…์ŠคํŠธ ์—†์ด๋Š” ์˜ฌ๋ฐ”๋ฅธ ํ–‰๋™์„ ์˜ˆ์ธกํ•˜๊ธฐ ์–ด๋ ต์Šต๋‹ˆ๋‹ค.

๊ฒฐ๋ก : ๋‹น์‹ ๋งŒ์˜ ์›Œํฌํ”Œ๋กœ๋ฅผ ์Šคํ‚ฌ๋กœ ์ฝ”๋”ฉํ•˜๋ผ

๋กœ์Šค ๋งˆ์ดํฌ์˜ ํ†ต์ฐฐ์€ AI ์—์ด์ „ํŠธ ํ™œ์šฉ์— ๋Œ€ํ•œ ์šฐ๋ฆฌ์˜ ์ ‘๊ทผ ๋ฐฉ์‹์„ ๊ทผ๋ณธ์ ์œผ๋กœ ๋ฐ”๊ฟ”์•ผ ํ•จ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค. ๋ชจ๋ธ์€ ๋›ฐ์–ด๋‚˜์ง€๋งŒ, โ€˜๋œ์–ด๋‚ผ์ˆ˜๋ก ๋” ํšจ๊ณผ์ โ€™์ด๋ผ๋Š” ๋ฏธ๋‹ˆ๋ฉ€๋ฆฌ์ฆ˜ ์ ‘๊ทผ ๋ฐฉ์‹์„ ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค. ๋ณต์žกํ•œ ์‹œ์Šคํ…œ์ด๋‚˜ ๋ฏธ๋ฆฌ ๋งŒ๋“ค์–ด์ง„ ์Šคํ‚ฌ์„ ๋งน๋ชฉ์ ์œผ๋กœ ๋”ฐ๋ฅด๊ธฐ๋ณด๋‹ค๋Š”, ๋‹น์‹  ๊ณ ์œ ์˜ ์›Œํฌํ”Œ๋กœ, ์ทจํ–ฅ, ์ „๋žต์„ ์—์ด์ „ํŠธ์™€ ํ•จ๊ป˜ ๋งŒ๋“ค์–ด๊ฐ€๊ณ , ์ด๋ฅผ ์Šคํ‚ฌ๋กœ ์ฝ”๋”ฉํ•˜๋Š” ๊ฒƒ์ด ์ค‘์š”ํ•ฉ๋‹ˆ๋‹ค.

๊ฒฐ๊ตญ ์ค‘์š”ํ•œ ๊ฒƒ์€ ๋ชจ๋ธ ์ž์ฒด์˜ ์„ฑ๋Šฅ์„ ๋„˜์–ด, ๋ชจ๋ธ์„ ๋‘˜๋Ÿฌ์‹ผ ํ•˜๋„ค์Šค(harness)์™€ ๋„๊ตฌ, ๊ทธ๋ฆฌ๊ณ  ๋‹น์‹ ์ด ์ œ๊ณตํ•˜๋Š” ์ปจํ…์ŠคํŠธ์ž…๋‹ˆ๋‹ค. ๋‹น์‹ ์˜ ๋น„์ฆˆ๋‹ˆ์Šค, ๋‹น์‹ ์˜ ์ž‘์—… ๋ฐฉ์‹์— ํŠนํ™”๋œ ์Šคํ‚ฌ์„ ์ง์ ‘ ๊ตฌ์ถ•ํ•˜๊ณ  ๋Š์ž„์—†์ด ๊ฐœ์„ ํ•ด๋‚˜๊ฐ€๋Š” ๊ฒƒ์ด AI ์‹œ๋Œ€์— ์ƒ์‚ฐ์„ฑ์„ ๊ทน๋Œ€ํ™”ํ•˜๊ณ  ๊ฒฝ์Ÿ๋ ฅ์„ ํ™•๋ณดํ•˜๋Š” ํ•ต์‹ฌ ์—ด์‡ ๊ฐ€ ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค.

๋กœ์Šค ๋งˆ์ดํฌ๋Š” ์ด ๊ฐ•์—ฐ์ด ๋‹จ์ง€ ์กฐํšŒ์ˆ˜๋ฅผ ์œ„ํ•œ ๊ฒƒ์ด ์•„๋‹ˆ๋ผ, ๋‹จ ํ•œ ๋ช…์ด๋ผ๋„ ์ด ์ •๋ณด๋ฅผ ํ†ตํ•ด ์‚ถ๊ณผ ๋น„์ฆˆ๋‹ˆ์Šค๋ฅผ ๋ณ€ํ™”์‹œํ‚ค๋Š” โ€˜์˜ํ–ฅ๋ ฅโ€™์— ์žˆ๋‹ค๊ณ  ๋งํ•˜๋ฉฐ ๊ฐ•์—ฐ์„ ๋งˆ๋ฌด๋ฆฌํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ์˜ ๋ฉ”์‹œ์ง€์ฒ˜๋Ÿผ, AI ์—์ด์ „ํŠธ์˜ ์ž ์žฌ๋ ฅ์„ ์ง„์ •์œผ๋กœ ํ•ด๋ฐฉ์‹œํ‚ค๋Š” ๊ฒƒ์€ ๊ธฐ์ˆ ์˜ ๋ณต์žก์„ฑ๋ณด๋‹ค๋Š” ์ธ๊ฐ„์˜ ์ง€ํ˜œ์™€ ์ธ๋‚ด์‹ฌ์— ๋‹ฌ๋ ค ์žˆ์Šต๋‹ˆ๋‹ค.


โ€œI built a custom Slack inbox. It was easier than you think. | Yash Tekriwal (Clay)โ€ โ€” How I AI ๊ธฐ๋ฐ˜ ๊ธฐ์‚ฌ ์›๋ณธ ์˜์ƒ ๋ณด๊ธฐ

AI๊ฐ€ ๋งŒ๋“  ๋‚˜๋งŒ์˜ ์Šฌ๋ž™ ๋น„์„œ: ์•Œ๋ฆผ ์ง€์˜ฅ์—์„œ ๋ฒ—์–ด๋‚˜๋Š” ๋ฒ•

๋งค์ผ ์Ÿ์•„์ง€๋Š” ์ˆ˜๋งŽ์€ ๋””์ง€ํ„ธ ์•Œ๋ฆผ ์†์—์„œ ์šฐ๋ฆฌ๋Š” ์ข…์ข… ์ค‘์š”ํ•œ ๋ฉ”์‹œ์ง€๋ฅผ ๋†“์น˜๊ฑฐ๋‚˜ ๋ถˆํ•„์š”ํ•œ ์ •๋ณด์— ์••๋„๋‹นํ•˜๊ณค ํ•ฉ๋‹ˆ๋‹ค. ํŠนํžˆ ์—…๋ฌด์šฉ ํ˜‘์—… ํˆด์ธ ์Šฌ๋ž™(Slack)์€ ๊ทธ ํŽธ๋ฆฌํ•จ๋งŒํผ์ด๋‚˜ ์—„์ฒญ๋‚œ ์–‘์˜ ์•Œ๋ฆผ์œผ๋กœ ์ธํ•ด ์‚ฌ์šฉ์ž๋“ค์„ โ€˜์Šฌ๋ž™ ์ง€์˜ฅโ€™์œผ๋กœ ๋ชฐ์•„๋„ฃ๊ธฐ๋„ ํ•ฉ๋‹ˆ๋‹ค. ํด๋ ˆ์ด(Clay)์˜ ๊ต์œก ์ฑ…์ž„์ž์ด์ž ๊ทน๋Œ€ํ™”๋œ ์ตœ์ ํ™” ์ „๋ฌธ๊ฐ€(hyper optimizer)์ธ ์•ผ์‹œ ํ…Œํฌ๋ฆฌ์›”(Yash Tekriwal)์€ ์ด๋Ÿฌํ•œ ์Šฌ๋ž™ ์•Œ๋ฆผ ๊ณผ๋ถ€ํ•˜ ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•ด AI๋ฅผ ํ™œ์šฉํ•œ ํ˜์‹ ์ ์ธ ๊ฐœ์ธ ๋งž์ถคํ˜• ์†”๋ฃจ์…˜์„ ๊ตฌ์ถ•ํ–ˆ์Šต๋‹ˆ๋‹ค.

์ด ๊ธฐ์‚ฌ์—์„œ๋Š” ์•ผ์‹œ๊ฐ€ ์–ด๋–ป๊ฒŒ AI ์ฝ”๋”ฉ ์—์ด์ „ํŠธ(AI coding agent)์™€ ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ(Perplexity Computer)๋ฅผ ํ™œ์šฉํ•˜์—ฌ ์ž์‹ ๋งŒ์˜ ์Šฌ๋ž™ ์•Œ๋ฆผ ๊ด€๋ฆฌ ์‹œ์Šคํ…œ์„ ๋งŒ๋“ค์—ˆ๋Š”์ง€, ๊ทธ๋ฆฌ๊ณ  ์ด๋Ÿฌํ•œ ์‹œ๋„๊ฐ€ ์‚ฌ์Šค(SaaS)์˜ ๋ฏธ๋ž˜์™€ ๊ฐœ์ธ์˜ ์ƒ์‚ฐ์„ฑ์— ์–ด๋–ค ์˜๋ฏธ๋ฅผ ๊ฐ€์ง€๋Š”์ง€ ์‹ฌ์ธต์ ์œผ๋กœ ๋‹ค๋ฃน๋‹ˆ๋‹ค.

1. ๋””์ง€ํ„ธ ์ •๋ณด ๊ณผ๋ถ€ํ•˜: ์Šฌ๋ž™ ์ง€์˜ฅ์—์„œ ๋ฒ—์–ด๋‚˜๊ธฐ

์•ผ์‹œ๋Š” ๋งค์ผ ์•„์นจ 100๊ฐœ์—์„œ 150๊ฐœ์— ๋‹ฌํ•˜๋Š” ์ƒˆ๋กœ์šด ์Šฌ๋ž™ ์•Œ๋ฆผ๊ณผ ํ•จ๊ป˜ ์ž ์—์„œ ๊นฌ๋‹ค๊ณ  ๋งํ•ฉ๋‹ˆ๋‹ค. ๋‹จ์ˆœํžˆ ์ฝ์ง€ ์•Š์€ ์ฑ„๋„ ๋ฉ”์‹œ์ง€๊ฐ€ ์•„๋‹ˆ๋ผ, ๋ˆ„๊ตฐ๊ฐ€๊ฐ€ ์ž์‹ ์„ ์ง์ ‘ ํƒœ๊ทธํ–ˆ๊ฑฐ๋‚˜ DM์„ ๋ณด๋‚ธ ์•Œ๋ฆผ๋“ค์ž…๋‹ˆ๋‹ค. ๋ฌธ์ œ๋Š” ์ด ์•Œ๋ฆผ๋“ค ์ค‘ 6080%๊ฐ€ โ€˜์ฐธ๊ณ ์šฉ(FYI: For Your Information)โ€™ ๋ฉ”์‹œ์ง€๋ผ๋Š” ์ ์ž…๋‹ˆ๋‹ค. ์ฆ‰, ๊ทธ์—๊ฒŒ ๋ถˆ์•ˆ๊ฐ์„ ์ฃผ๋Š” 100150๊ฐœ์˜ ์•Œ๋ฆผ ์ค‘ ์‹ค์ œ๋กœ ์ฆ‰๊ฐ์ ์ธ ์กฐ์น˜๊ฐ€ ํ•„์š”ํ•œ ๊ฒƒ์€ 30~40๊ฐœ์— ๋ถˆ๊ณผํ•œ ๊ฒƒ์ž…๋‹ˆ๋‹ค.

์•ผ์‹œ๋Š” ์ด ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•ด โ€˜๋ชจ๋“  ์•Œ๋ฆผ์ด ๋™๋“ฑํ•˜๊ฒŒ ์ƒ์„ฑ๋˜์ง€ ์•Š๋Š”๋‹คโ€™๋Š” ํ•ต์‹ฌ ํ†ต์ฐฐ๋ ฅ์„ ์–ป์—ˆ์Šต๋‹ˆ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด, ํŒŸ์บ์ŠคํŠธ ๋…น์Œ ์ผ์ •์„ ์กฐ์œจํ•˜๋Š” ์ค‘์š”ํ•œ DM๊ณผ ๋™๋ฃŒ๊ฐ€ ๋ฐ˜๋ ค๊ฒฌ ์‚ฌ์ง„์„ ์˜ฌ๋ฆฐ ์ฑ„๋„์˜ ์œ ์พŒํ•œ ๋Œ“๊ธ€ ์•Œ๋ฆผ์€ ๋™๋“ฑํ•˜๊ฒŒ ์ทจ๊ธ‰๋  ์ˆ˜ ์—†์Šต๋‹ˆ๋‹ค. ๊ทธ๋Š” ์ด๋Ÿฌํ•œ ์•Œ๋ฆผ๋“ค์„ ๋”์šฑ ์ฒด๊ณ„์ ์œผ๋กœ ๋ถ„๋ฅ˜ํ•˜๊ณ  ์šฐ์„ ์ˆœ์œ„๋ฅผ ์ •ํ•  ์ˆ˜ ์žˆ๋Š” ์‹œ์Šคํ…œ์ด ํ•„์š”ํ•˜๋‹ค๊ณ  ๋А๊ผˆ์Šต๋‹ˆ๋‹ค.

2. AI ๋น„์„œ, ์Šฌ๋ž™ ์•Œ๋ฆผ ๋ถ„๋ฅ˜ ์‹œ์Šคํ…œ์„ ๊ตฌ์ถ•ํ•˜๋‹ค: ์˜คํ”ˆํด๋กœ(OpenClaw)์˜ ์ฒซ ์‹œ๋„

์•ผ์‹œ๋Š” AI๋ฅผ ํ™œ์šฉํ•˜๋Š” ๋‘ ๊ฐ€์ง€ ๋ฐฉ๋ฒ•์„ ๋ช…ํ™•ํžˆ ๊ตฌ๋ถ„ํ•ฉ๋‹ˆ๋‹ค. ํ•˜๋‚˜๋Š” โ€˜๋ถ„๋ฅ˜ํ•˜๊ฑฐ๋‚˜ ์š”์•ฝํ•˜๋Š” ๋“ฑ ํŠน์ • ์ž‘์—…์„ ์ˆ˜ํ–‰ํ•˜๊ฒŒ ํ•˜๋Š” ๊ฒƒโ€™์ด๊ณ , ๋‹ค๋ฅธ ํ•˜๋‚˜๋Š” โ€˜์ด์ „์—๋Š” ํ›จ์”ฌ ์–ด๋ ค์› ์„ ๋„๊ตฌ๋ฅผ ์ง์ ‘ ๊ตฌ์ถ•ํ•˜๊ฒŒ ํ•˜๋Š” ๊ฒƒโ€™์ž…๋‹ˆ๋‹ค. ๊ทธ์˜ ์Šฌ๋ž™ ๋ฌธ์ œ ํ•ด๊ฒฐ์€ ํ›„์ž์— ํ•ด๋‹นํ–ˆ์Šต๋‹ˆ๋‹ค.

๊ทธ๋Š” ๋จผ์ € ์˜คํ”ˆํด๋กœ(OpenClaw)๋ผ๋Š” AI ์ฝ”๋”ฉ ์—์ด์ „ํŠธ๋ฅผ ๋””์Šค์ฝ”๋“œ(Discord)์—์„œ ํ™œ์šฉํ•˜์—ฌ ์Šฌ๋ž™ ์•Œ๋ฆผ ๋ถ„๋ฅ˜ ์‹œ์Šคํ…œ์˜ ํ”„๋กœํ† ํƒ€์ž…์„ ๊ตฌ์ถ•ํ–ˆ์Šต๋‹ˆ๋‹ค. ๋””์Šค์ฝ”๋“œ๋ฅผ ์„ ํƒํ•œ ์ด์œ ๋Š” ์Šฌ๋ž™๊ณผ ์œ ์‚ฌํ•œ ์“ฐ๋ ˆ๋“œ(threading) ๊ธฐ๋Šฅ๊ณผ ๊ฐ•๋ ฅํ•œ ๊ฒ€์ƒ‰ ๊ธฐ๋Šฅ ๋•๋ถ„์— ๋Œ€ํ™”์˜ ๋งฅ๋ฝ์„ ์ •๋ฆฌํ•˜๊ณ  ์‰ฝ๊ฒŒ ์ฐพ์•„๋ณผ ์ˆ˜ ์žˆ์—ˆ๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.

๊ตฌ์ถ• ๊ณผ์ •:

  1. ๋ฌธ์ œ ์ •์˜: ์Šฌ๋ž™์˜ ๋ชจ๋“  ๋ฉ”์‹œ์ง€๋ฅผ ๊ฐ€์ ธ์˜ค๋Š” ๊ฒƒ์€ ์••๋„์ ์ด๋ฏ€๋กœ, ํ•„์š”ํ•œ ๋งฅ๋ฝ๋งŒ ๊ฐ€์ง„ ๋ฉ”์‹œ์ง€๋ฅผ ์„ ๋ณ„ํ•ด์•ผ ํ–ˆ์Šต๋‹ˆ๋‹ค. ์Šฌ๋ž™์˜ ์•Œ๋ฆผ ์‹œ์Šคํ…œ์ด โ€˜์ฝ์ง€ ์•Š์Œโ€™ ์ƒํƒœ, ์ˆซ์ž ํ‘œ๊ธฐ ๋ฐฉ์‹ ๋“ฑ์„ ๋งค์šฐ ์˜๋„์ ์œผ๋กœ ์„ค๊ณ„ํ•œ๋‹ค๋Š” ์ ์„ ํŒŒ์•…ํ–ˆ์Šต๋‹ˆ๋‹ค.
  2. ๋ฐ์ดํ„ฐ ๋ถ„์„: ์•ผ์‹œ๋Š” ์˜คํ”ˆํด๋กœ์™€ ์ˆ˜๋งŽ์€ ๋Œ€ํ™”๋ฅผ ์ฃผ๊ณ ๋ฐ›์œผ๋ฉฐ ์Šฌ๋ž™์˜ ํƒ€์ž„์Šคํƒฌํ”„(timestamp) ๋ฐ์ดํ„ฐ๋ฅผ ๋ถ„์„ํ–ˆ์Šต๋‹ˆ๋‹ค. ์–ธ์ œ ๋งˆ์ง€๋ง‰์œผ๋กœ ๋ฉ”์‹œ์ง€๋ฅผ ํ™•์ธํ–ˆ๋Š”์ง€, ์“ฐ๋ ˆ๋“œ ๋‚ด์—์„œ ๊ฐ€์žฅ ์ตœ๊ทผ ๋ฉ”์‹œ์ง€๋งŒ ํ•„์š”ํ•œ์ง€ ๋“ฑ ๋ณต์žกํ•œ ๋กœ์ง์„ AI์™€ ํ•จ๊ป˜ ์—ญ์„ค๊ณ„(reverse-engineering)ํ–ˆ์Šต๋‹ˆ๋‹ค.
  3. ์ฝ”๋“œ ์ƒ์„ฑ: ์˜คํ”ˆํด๋กœ๋Š” ์ด ๋ฐ์ดํ„ฐ๋ฅผ ๊ธฐ๋ฐ˜์œผ๋กœ ์•Œ๋ฆผ ์š”์•ฝ(digest)์„ ์ƒ์„ฑํ•˜๋Š” ์ฝ”๋“œ๋ฅผ ์ž‘์„ฑํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด ์š”์•ฝ์€ ๋‹ค์ด๋ ‰ํŠธ ๋ฉ˜์…˜(@mentions), DM, ๊ทธ๋ฃน ๋ฉ˜์…˜, ์“ฐ๋ ˆ๋“œ ๋“ฑ ๋„ค ๊ฐ€์ง€ ์ฃผ์š” ๋ฒ”์ฃผ๋กœ ๋‚˜๋‰˜๊ณ , ๊ฐ ๋ฒ”์ฃผ ๋‚ด์—์„œ ๋‹ค์‹œ โ€˜์กฐ์น˜ ํ•„์š”โ€™, โ€˜์ฝ๊ธฐ ํ•„์š”โ€™, โ€˜์ฐธ๊ณ ์šฉ(FYI)โ€˜์œผ๋กœ ์„ธ๋ถ„ํ™”๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

๊ฒฐ๊ณผ:

์ด ๊ณผ์ •์„ ํ†ตํ•ด ์•ผ์‹œ๋Š” ์Šฌ๋ž™ ๋‚ด์— โ€˜์ž๋น„์Šค(Jarvis) ๋‹ค์ด์ œ์ŠคํŠธโ€™๋ผ๋Š” ์ฑ„๋„์„ ๋งŒ๋“ค์—ˆ์Šต๋‹ˆ๋‹ค. ์ด๊ณณ์—๋Š” ํ•˜๋ฃจ 100๊ฐœ๊ฐ€ ๋„˜๋Š” ์•Œ๋ฆผ์ด ์ด์ œ๋Š” ์ž˜ ์ •๋ฆฌ๋œ ํ…์ŠคํŠธ ๋ชฉ๋ก ํ˜•ํƒœ๋กœ ๊ทธ๋ฃนํ™”๋˜์–ด ๋‚˜ํƒ€๋‚ฌ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Š” ์ด์ œ ๊ฐ ์•Œ๋ฆผ์„ ํด๋ฆญํ•˜์—ฌ ํ•ด๋‹น ์“ฐ๋ ˆ๋“œ๋กœ ์ด๋™ํ•˜๊ณ , ์กฐ์น˜๋ฅผ ์ทจํ•œ ํ›„ ๋‹ค์‹œ ๋Œ์•„์˜ค๋Š” ๋ฐฉ์‹์œผ๋กœ ์—…๋ฌด๋ฅผ ์ฒ˜๋ฆฌํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

ํ•˜์ง€๋งŒ ์ด ํ…์ŠคํŠธ ๊ธฐ๋ฐ˜์˜ ์š”์•ฝ์€ ์—ฌ์ „ํžˆ ํ•œ๊ณ„๊ฐ€ ์žˆ์—ˆ์Šต๋‹ˆ๋‹ค. ์Šคํฌ๋กค์„ 4~5๋ฒˆ ๋‚ด๋ ค์•ผ ๋ชจ๋“  ์•Œ๋ฆผ์„ ๋ณผ ์ˆ˜ ์žˆ์—ˆ๊ณ , ์‹œ๊ฐ์ ์œผ๋กœ ์••๋„์ ์ด๋ฉฐ ์ •์‹ ์ ์œผ๋กœ ์†Œ๋ชจ๊ฐ€ ์ปธ์Šต๋‹ˆ๋‹ค. ์•ผ์‹œ๋Š” ์—ฌ๊ธฐ์„œ ํ•œ ๋‹จ๊ณ„ ๋” ๋‚˜์•„๊ฐ€ โ€˜์‹ค์ œ ์†Œํ”„ํŠธ์›จ์–ดโ€™ ํ˜•ํƒœ์˜ ์‚ฌ์šฉ์ž ์ธํ„ฐํŽ˜์ด์Šค(UI)๋ฅผ ๊ฟˆ๊พธ๊ธฐ ์‹œ์ž‘ํ–ˆ์Šต๋‹ˆ๋‹ค.

3. ์ง๊ด€์ ์ธ ๊ฐœ์ธ ๋งž์ถค ๋Œ€์‹œ๋ณด๋“œ: ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ(Perplexity Computer)์˜ ๋งˆ๋ฒ•

์•ผ์‹œ์˜ ๋‹ค์Œ ๋ชฉํ‘œ๋Š” โ€˜์Šˆํผํœด๋จผ(Superhuman)โ€™ ์ด๋ฉ”์ผ ์•ฑ์ฒ˜๋Ÿผ ๊น”๋”ํ•˜๊ณ  ์ง๊ด€์ ์ธ ๋Œ€์‹œ๋ณด๋“œ๋ฅผ ๊ตฌ์ถ•ํ•˜๋Š” ๊ฒƒ์ด์—ˆ์Šต๋‹ˆ๋‹ค. ์—ฌ๊ธฐ์— ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ(Perplexity Computer)๊ฐ€ ๋“ฑ์žฅํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Š” ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ๊ธฐ์กด์˜ ํ…์ŠคํŠธ ์š”์•ฝ์„ ์‹œ๊ฐ์ ์ธ ์นธ๋ฐ˜(Kanban) ์Šคํƒ€์ผ ๋ณด๋“œ(board)๋กœ ์ „ํ™˜ํ–ˆ์Šต๋‹ˆ๋‹ค.

ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ๋กœ ๊ตฌ์ถ•ํ•œ ์Šฌ๋ž™ ๊ด€๋ฆฌ ๋Œ€์‹œ๋ณด๋“œ:

  • ์‹œ๊ฐ์  ๋ถ„๋ฅ˜: ๋Œ€์‹œ๋ณด๋“œ๋Š” ์„ธ ๊ฐœ์˜ ์ฃผ์š” ์—ด๋กœ ๊ตฌ์„ฑ๋ฉ๋‹ˆ๋‹ค.
    • ๋นจ๊ฐ„์ƒ‰ (์™ผ์ชฝ): ์กฐ์น˜ ํ•„์š”(Action Required, Urgent) โ€“ ์•ผ์‹œ๊ฐ€ ์ฆ‰์‹œ ์ฒ˜๋ฆฌํ•ด์•ผ ํ•  ์•Œ๋ฆผ.
    • ๋…ธ๋ž€์ƒ‰ (์ค‘์•™): ์ฝ๊ธฐ ํ•„์š”(Need to Read) โ€“ ๋‚ด์šฉ์„ ์ดํ•ดํ•ด์•ผ ํ•˜์ง€๋งŒ ์ฆ‰๊ฐ์ ์ธ ์‘๋‹ต์ด ํ•„์š” ์—†๋Š” ์•Œ๋ฆผ.
    • ์ดˆ๋ก์ƒ‰ (์˜ค๋ฅธ์ชฝ): ์ฐธ๊ณ ์šฉ(FYI) โ€“ ์ •๋ณด์„ฑ ๋ฉ”์‹œ์ง€๋กœ, ๋ถ€๋‹ด ์—†์ด ํ™•์ธํ•  ์ˆ˜ ์žˆ๋Š” ์•Œ๋ฆผ.
  • ๋งž์ถคํ˜• ์ •๋ ฌ: ์™ผ์ชฝ ์‚ฌ์ด๋“œ๋ฐ”์—์„œ DM, ๊ทธ๋ฃน ๋ฉ˜์…˜ ๋“ฑ ์•Œ๋ฆผ ์œ ํ˜•๋ณ„๋กœ ์ •๋ ฌํ•˜์—ฌ ํ™•์ธํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • โ€œ๋ชจ๋‘ ์•„์นด์ด๋ธŒ(Archive All)โ€ ๊ธฐ๋Šฅ: ๊ฐ€์žฅ ํ˜์‹ ์ ์ธ ๊ธฐ๋Šฅ์€ โ€˜์ฐธ๊ณ ์šฉ(FYI)โ€™ ์—ด์— ์žˆ๋Š” โ€œ๋ชจ๋‘ ์•„์นด์ด๋ธŒโ€ ๋ฒ„ํŠผ์ž…๋‹ˆ๋‹ค. ์ด ๋ฒ„ํŠผ์„ ํด๋ฆญํ•˜๋ฉด ํ•ด๋‹น ์•Œ๋ฆผ๋“ค์ด ๋Œ€์‹œ๋ณด๋“œ์—์„œ ์‚ฌ๋ผ์ง€๋Š” ๋™์‹œ์—, ์Šฌ๋ž™์—์„œ๋„ ํ•ด๋‹น ์•Œ๋ฆผ์ด ๋ชจ๋‘ ์‚ฌ๋ผ์ ธ ๊นจ๋—ํ•œ ์ƒํƒœ๋ฅผ ์œ ์ง€ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์Šฌ๋ž™์˜ โ€˜๋ชจ๋‘ ์ฝ์Œ์œผ๋กœ ํ‘œ์‹œโ€™ ๊ธฐ๋Šฅ๋ณด๋‹ค ํ›จ์”ฌ ์ •๊ตํ•˜์—ฌ, ์•ผ์‹œ์˜ ์—…๋ฌด ํ๋ฆ„์— ์™„๋ฒฝํ•˜๊ฒŒ ๋งž์ถคํ™”๋œ ๊ฒฝํ—˜์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.

์™œ ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ์ธ๊ฐ€?

์•ผ์‹œ๋Š” ์˜คํ”ˆํด๋กœ์™€ ๊ฐ™์€ ๋‹ค๋ฅธ AI ๋„๊ตฌ๋“ค๋ณด๋‹ค ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ์— ๋งค๋ฃŒ๋œ ์„ธ ๊ฐ€์ง€ ์ฃผ์š” ์ด์œ ๋ฅผ ์„ค๋ช…ํ•ฉ๋‹ˆ๋‹ค.

  1. ๋ณ‘๋ ฌ ์ฒ˜๋ฆฌ(Concurrent Runs) ๋ฐ ์žฅ๊ธฐ ์‹คํ–‰ ์ž‘์—…: ํด๋กœ๋“œ(Claude)๋‚˜ ์ฝ”๋ฑ์Šค(CodeX)๋Š” ๊ธฐ๋ณธ์ ์œผ๋กœ ์ฑ„ํŒ… ๊ธฐ๋ฐ˜์˜ ์ˆœ์ฐจ์  ๋Œ€ํ™” ๋ฐฉ์‹์ž…๋‹ˆ๋‹ค. ํ•˜์ง€๋งŒ ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ๋Š” ์—ฌ๋Ÿฌ ์ž‘์—…์„ ๋™์‹œ์— ์‹œ์ž‘ํ•˜๊ณ  ๋ณ‘๋ ฌ๋กœ ์‹คํ–‰ํ•  ์ˆ˜ ์žˆ์–ด, ํ›จ์”ฌ ํšจ์œจ์ ์œผ๋กœ ๊ฐœ๋ฐœํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  2. ํด๋ผ์šฐ๋“œ ๊ธฐ๋ฐ˜์˜ ๋„ค์ดํ‹ฐ๋ธŒ ์—ฐ๊ฒฐ์„ฑ(Cloud-Native Connectors): ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ๋Š” ํด๋ผ์šฐ๋“œ์— ์žˆ์–ด ๊ตฌ๊ธ€ ๋“œ๋ผ์ด๋ธŒ(Google Drive), ์ง€๋ฉ”์ผ(Gmail), ๋…ธ์…˜(Notion), ์•„์‚ฌ๋‚˜(Asana), ์Šฌ๋ž™(Slack), ์คŒ(Zoom) ๋“ฑ ๋‹ค์–‘ํ•œ ์•ฑ๊ณผ ๊ธฐ๋ณธ์ ์œผ๋กœ ์—ฐ๊ฒฐ๋ฉ๋‹ˆ๋‹ค. ์•ผ์‹œ๋Š” ์˜คํ”ˆํด๋กœ์ฒ˜๋Ÿผ ๊ฐ ์†Œํ”„ํŠธ์›จ์–ด์— ์ ‘๊ทผํ•˜๊ธฐ ์œ„ํ•ด ์Šคํ‚ฌ(skills)์„ ๋”ฐ๋กœ ์ œ๊ณตํ•  ํ•„์š” ์—†์ด, ๋กœ๊ทธ์ธ๋งŒ ํ•˜๋ฉด ์ปดํ“จํ„ฐ๊ฐ€ ์•Œ์•„์„œ ์—ฐ๋™ํ•˜์—ฌ ๋ฐ์ดํ„ฐ๋ฅผ ๊ฐ€์ ธ์˜ค๊ณ  ์ž‘์—…์„ ์ˆ˜ํ–‰ํ•ฉ๋‹ˆ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด, ๋…ธ์…˜ AI(Notion AI)๋กœ ๊ธฐ๋ก๋œ ๋ชจ๋“  ํšŒ์˜๋ก์—์„œ ์กฐ์น˜ ํ•ญ๋ชฉ์„ ์ถ”์ถœํ•˜๊ณ , ์ค‘์š”ํ•œ ๊ฒƒ์€ ์•„์‚ฌ๋‚˜(Asana)์— ๋“ฑ๋กํ•˜๋ฉฐ, ๊ฐ„๋‹จํ•œ ๋ฉ”์‹œ์ง€๋‚˜ ์•Œ๋ฆผ, ์ด๋ฉ”์ผ์—๋Š” ์‘๋‹ต ์ดˆ์•ˆ์„ ์ž‘์„ฑํ•˜๋„๋ก ์ง€์‹œํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด์ฒ˜๋Ÿผ ๋Š๊น€ ์—†๋Š”(fluid) ์—ฐ๋™์€ ์ž‘์—… ์†๋„๋ฅผ ํ˜์‹ ์ ์œผ๋กœ ๋†’์—ฌ์ค๋‹ˆ๋‹ค.
  3. ์•™์ƒ๋ธ” ์˜ค์ผ€์ŠคํŠธ๋ ˆ์ด์…˜(Ensemble Orchestration)์„ ํ†ตํ•œ ์ง€๋Šฅ์  ๊ตฌ์ถ•: ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ๋Š” ์†Œ๋„คํŠธ(Sonnet), ์ œ๋ฏธ๋‹ˆ(Gemini), ์˜คํ‘ธ์Šค(Opus) ๋“ฑ ๋‹ค์–‘ํ•œ AI ๋ชจ๋ธ์„ ์ž‘์—…์˜ ๊ฐ ๋ถ€๋ถ„์— ๋งž๊ฒŒ ์œ ์—ฐํ•˜๊ฒŒ ํ™œ์šฉํ•ฉ๋‹ˆ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด, ๋ฐ์ดํ„ฐ ๊ฒ€์ƒ‰์—๋Š” ์†Œ๋„คํŠธ๋ฅผ, ๊ณ„ํš ์ˆ˜๋ฆฝ์—๋Š” ์ œ๋ฏธ๋‹ˆ๋ฅผ, ์‹ค์ œ ์ฝ”๋”ฉ ๋ฐ ๋ณต์žกํ•œ ์ถ”๋ก ์—๋Š” ์˜คํ‘ธ์Šค๋ฅผ ์‚ฌ์šฉํ•˜๋Š” ์‹์ž…๋‹ˆ๋‹ค. ์ด โ€˜์•™์ƒ๋ธ” ์˜ค์ผ€์ŠคํŠธ๋ ˆ์ด์…˜โ€™ ๋•๋ถ„์— ์ธ๊ฐ„์ด ์ผ์ผ์ด ํ”ผ๋“œ๋ฐฑ์„ ์ฃผ๋ฉฐ โ€œ๋‹ค์‹œ ์‹œ๋„ํ•ด๋ดโ€๋ผ๊ณ  ๋งํ•  ํ•„์š” ์—†์ด, ์Šค์Šค๋กœ ๋ฌธ์ œ ํ•ด๊ฒฐ ๋ฐ ํ…Œ์ŠคํŠธ ๊ณผ์ •์„ ๊ฑฐ์ณ ์ •ํ™•ํ•œ ๊ฒฐ๊ณผ๋ฌผ์„ ๋งŒ๋“ค์–ด๋ƒ…๋‹ˆ๋‹ค. ๋˜ํ•œ, ๊ฐœ๋ฐœ๋œ ์•ฑ์ด ์ž๋™์œผ๋กœ ๋ฐฐํฌ๋˜์–ด ์›น์—์„œ ๋ฐ”๋กœ ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ๋‹ค๋Š” ์ ๋„ ํฐ ์žฅ์ ์ž…๋‹ˆ๋‹ค.

4. SaaS์˜ ๋ฏธ๋ž˜: ๋งž์ถคํ˜• ๊ฒฝํ—˜์˜ ํญ๋ฐœ์  ์„ฑ์žฅ

์•ผ์‹œ์™€ ํด๋ ˆ์–ด(Claire Vo)๋Š” โ€œ์‚ฌ์Šค(SaaS)๋Š” ์ฃฝ์—ˆ๋Š”๊ฐ€?โ€๋ผ๋Š” ์งˆ๋ฌธ์— ๋Œ€ํ•ด โ€œ์•„๋‹ˆ๋‹ค, ์ง„ํ™”ํ•˜๊ณ  ์žˆ๋‹คโ€๊ณ  ๋‹ตํ•ฉ๋‹ˆ๋‹ค. ์Šฌ๋ž™ ์ž์ฒด๋Š” ๋ฉ”์‹œ์ง€๋ฅผ ๋ณด๋‚ด๋Š” ๋ฐ ์—ฌ์ „ํžˆ ํ›Œ๋ฅญํ•˜์ง€๋งŒ, ๋ฉ”์‹œ์ง€๋ฅผ ์ฝ๋Š” ๋ฐฉ์‹์€ ๊ฐœ์„ ์˜ ์—ฌ์ง€๊ฐ€ ์žˆ์Šต๋‹ˆ๋‹ค. AI๋ฅผ ํ†ตํ•ด ์‚ฌ์šฉ์ž๋Š” ๊ธฐ์กด ์‚ฌ์Šค ์œ„์— ์ž์‹ ๋งŒ์˜ ๋งž์ถคํ˜• ๋ ˆ์ด์–ด๋ฅผ ๊ตฌ์ถ•ํ•˜์—ฌ ๊ฐœ์ธ์˜ ์ด์ƒ์ ์ธ ์—…๋ฌด ํ๋ฆ„์— ์™„๋ฒฝํ•˜๊ฒŒ ๋งž์ถœ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ๊ธฐ์กด ์†Œํ”„ํŠธ์›จ์–ด์˜ โ€˜๊ฒฐํ•จโ€™ ๋•Œ๋ฌธ์ด ์•„๋‹ˆ๋ผ, โ€˜๊ฐœ์ธ์˜ ์ตœ์ ํ™”๋œ ์›Œํฌํ”Œ๋กœ์šฐโ€™์™€ โ€˜์‚ฌ์Šค๊ฐ€ ์ œ๊ณตํ•˜๋Š” ๊ธฐ๋ณธ ๊ธฐ๋Šฅโ€™ ์‚ฌ์ด์˜ ๊ฐ„๊ทน์„ ๋ฉ”์šฐ๋Š” ํ–‰์œ„์ž…๋‹ˆ๋‹ค.

์ด๋Ÿฌํ•œ ๋ณ€ํ™”๋Š” โ€˜๋งˆ์ดํฌ๋กœ ์‚ฌ์Šค(Micro-SaaS)โ€˜์˜ ํญ๋ฐœ์ ์ธ ์„ฑ์žฅ์„ ์˜ˆ๊ณ ํ•ฉ๋‹ˆ๋‹ค. ์ด์ „์—๋Š” ๋น„์šฉ ๋ฌธ์ œ๋กœ ๊ฐœ๋ฐœํ•  ๊ฐ€์น˜๊ฐ€ ์—†์—ˆ๋˜ ํ‹ˆ์ƒˆ ์†Œํ”„ํŠธ์›จ์–ด๋“ค์ด ์ด์ œ๋Š” AI ๋•๋ถ„์— ์ €๋ ดํ•˜๊ฒŒ ๊ตฌ์ถ•๋  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์•ผ์‹œ๋Š” ์ž์‹ ์˜ ์Šฌ๋ž™ ๊ด€๋ฆฌ ์•ฑ์„ ๋ˆ„๊ตฐ๊ฐ€ ์ œํ’ˆ์œผ๋กœ ๋งŒ๋“ค์–ด ์›” 15๋‹ฌ๋Ÿฌ์— ํŒ๋งคํ•œ๋‹ค๋ฉด ๊ธฐ๊บผ์ด ๊ตฌ๋งคํ•˜์—ฌ ๋ฒ„๊ทธ ๋ฆฌํฌํŠธ๋ฅผ ์ œ์ถœํ•˜๊ฒ ๋‹ค๊ณ  ๋งํ•ฉ๋‹ˆ๋‹ค. ๋ฒค์ฒ˜ ํˆฌ์ž๋ฅผ ๋ฐ›๊ฑฐ๋‚˜ ์ˆ˜์‹ญ์–ต ๋‹ฌ๋Ÿฌ ๊ทœ๋ชจ๋กœ ์„ฑ์žฅํ•  ํ•„์š” ์—†์ด, ์ˆ˜์‹ญ๋งŒ ๋‹ฌ๋Ÿฌ ๊ทœ๋ชจ์˜ ์œ ์šฉํ•˜๊ณ  ์ˆ˜์ต์„ฑ ์žˆ๋Š” ์†Œํ”„ํŠธ์›จ์–ด๋“ค์ด ๋„˜์ณ๋‚  ๊ฒƒ์ด๋ผ๋Š” ์˜ˆ์ธก์ž…๋‹ˆ๋‹ค.

์•ผ์‹œ๋Š” ๋˜ํ•œ โ€˜์•ˆํ‹ฐ-ํˆฌ๋‘ ๋ฆฌ์ŠคํŠธ(Anti-to-do list)โ€˜๋ผ๋Š” ๊ฐœ๋…์„ ์ œ์‹œํ•ฉ๋‹ˆ๋‹ค. ์ด๋Š” โ€˜์ ˆ๋Œ€ ๋‹ค์‹œ๋Š” ์ˆ˜๋™์œผ๋กœ ํ•˜๊ณ  ์‹ถ์ง€ ์•Š์€ ์ผโ€™์˜ ๋ชฉ๋ก์œผ๋กœ, AI๋ฅผ ํ™œ์šฉํ•ด ์ด๋Ÿฌํ•œ ๋ฐ˜๋ณต์ ์ด๊ณ  ์„ฑ๊ฐ€์‹  ์ž‘์—…์„ ์ž๋™ํ™”ํ•˜๋Š” ๋ฐ ์‹œ๊ฐ„์„ ํˆฌ์žํ•˜๋Š” ๊ฒƒ์ด ๋งค์šฐ ๊ฐ€์น˜ ์žˆ๋‹ค๊ณ  ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค. ์ŠคํŒธ ์ด๋ฉ”์ผ ์ˆ˜๋™ ์‚ญ์ œ, ํšŒ์˜ ํ›„ ์•„์‚ฌ๋‚˜์— ์ˆ˜๋™์œผ๋กœ ์กฐ์น˜ ํ•ญ๋ชฉ ์ž…๋ ฅ, ๋น„์ •ํ˜•์ ์ธ ์Šฌ๋ž™ ์•Œ๋ฆผ ํ™•์ธ ๋“ฑ์ด ๊ทธ ์˜ˆ์‹œ๊ฐ€ ๋  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

5. ๊ฐœ์ธ ์ƒ์‚ฐ์„ฑ์„ ๋„˜์–ด, ํŒ€ ํ˜‘์—…๊ณผ ํ˜์‹ ์œผ๋กœ

์•ผ์‹œ์˜ AI ํ™œ์šฉ์€ ๊ฐœ์ธ ์ƒ์‚ฐ์„ฑ ํ–ฅ์ƒ์—๋งŒ ๊ทธ์น˜์ง€ ์•Š์Šต๋‹ˆ๋‹ค. ๊ทธ๋Š” ์Šฌ๋ž™, ์ด๋ฉ”์ผ, ๋‰ด์Šค ๋“ฑ ๋‹ค์–‘ํ•œ ์ •๋ณด ์ŠคํŠธ๋ฆผ์„ ํ†ตํ•ฉํ•˜์—ฌ ์ž์‹ ๋งŒ์˜ โ€˜๊ฐœ์ธ ์‚ฌ๋ น๋ถ€(command center)โ€™ ๋Œ€์‹œ๋ณด๋“œ๋ฅผ ๊ตฌ์ถ•ํ•˜๊ธฐ๋„ ํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด ๋˜ํ•œ ์นธ๋ฐ˜ ์Šคํƒ€์ผ๋กœ, ๋ฏธ์…˜ ํฌ๋ฆฌํ‹ฐ์ปฌํ•œ ์ •๋ณด๋งŒ ์„ ๋ณ„ํ•˜์—ฌ ํ‘œ์‹œํ•˜๊ณ  ์™ธ๋ถ€ ๋งํฌ๋กœ ๋ฐ”๋กœ ์ด๋™ํ•  ์ˆ˜ ์žˆ๋„๋ก ์„ค๊ณ„๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

๋” ๋‚˜์•„๊ฐ€, ๊ทธ์˜ ํŒ€์›์ธ ํฌ๋ฆฌ์Šค ๋ฐ(Chris Ming)์€ ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ๋ฅผ ํ™œ์šฉํ•˜์—ฌ ํด๋ ˆ์ด ์œ ๋‹ˆ๋ฒ„์‹œํ‹ฐ(Clay University) ์›น์‚ฌ์ดํŠธ์˜ ํ”„๋กœํ† ํƒ€์ž…์„ ๊ตฌ์ถ•ํ–ˆ์Šต๋‹ˆ๋‹ค. ํด๋ ˆ์ด ์œ ๋‹ˆ๋ฒ„์‹œํ‹ฐ๋Š” ํด๋ ˆ์ด ์ œํ’ˆ ํ•™์Šต์„ ์œ„ํ•œ ํ’๋ถ€ํ•œ ์ฝ˜ํ…์ธ ๋ฅผ ์ œ๊ณตํ•˜์ง€๋งŒ, ๋‹ค์–‘ํ•œ ์ง๋ฌด(RevOps, ๋งˆ์ผ€ํŒ…, GTM ์—”์ง€๋‹ˆ์–ด ๋“ฑ)์— ๋งž์ถฐ ๊ฐœ์ธํ™”๋œ ํ•™์Šต ๊ฒฝ๋กœ๋ฅผ ์ œ๊ณตํ•˜๋Š” ๋ฐ๋Š” ํ•œ๊ณ„๊ฐ€ ์žˆ์—ˆ์Šต๋‹ˆ๋‹ค.

ํฌ๋ฆฌ์Šค๋Š” ํผํ”Œ๋ ‰์‹œํ‹ฐ ์ปดํ“จํ„ฐ๊ฐ€ ๋ธŒ๋ผ์šฐ์ €๋ฅผ ํ†ตํ•ด ์›น์‚ฌ์ดํŠธ์— ์ ‘๊ทผํ•˜๊ณ  ์‹œ๊ฐ์  ๋งฅ๋ฝ์„ ์ดํ•ดํ•  ์ˆ˜ ์žˆ๋‹ค๋Š” ์ ์„ ํ™œ์šฉํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Š” ํ•œ ์‹œ๊ฐ„ ๋™์•ˆ ์ปดํ“จํ„ฐ์™€ ๋Œ€ํ™”ํ•˜๋ฉฐ, ๋กœ๊ทธ์ธ ์ƒํƒœ์™€ ์ง๋ฌด์— ๋”ฐ๋ผ SDR, BDR, RevOps ๋“ฑ ๊ฐ ํŽ˜๋ฅด์†Œ๋‚˜(persona)์—๊ฒŒ ์ ํ•ฉํ•œ ํ•™์Šต ๊ฒฝ๋กœ์™€ ๋‹ค์Œ ๊ฐ•์ขŒ, ์›Œํฌ์ˆ ๋“ฑ์„ ๋ณด์—ฌ์ฃผ๋Š” ๋งž์ถคํ˜• UI๋ฅผ ๋งŒ๋“ค์–ด๋ƒˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ๋””์ž์ธ ํŒ€์ด ์ƒˆ๋กœ์šด ๋””์ž์ธ ์‹œ์Šคํ…œ์„ ๊ตฌ์ƒํ•˜๊ณ  ๊ฐœ๋ฐœํ•˜๋Š” ๋ฐ ํ•„์š”ํ•œ ์‹œ๊ฐ์  โ€˜๋‹ค๋ฆฌ(bridge)โ€™ ์—ญํ• ์„ ํ•˜์—ฌ, ๋””์ž์ธ๊ณผ ๋‹ค๋ฅธ ์ดํ•ด๊ด€๊ณ„์ž๋“ค ๊ฐ„์˜ ์ปค๋ฎค๋‹ˆ์ผ€์ด์…˜ ๊ฐ„๊ทน์„ ํšจ๊ณผ์ ์œผ๋กœ ๋ฉ”์›Œ์ฃผ์—ˆ์Šต๋‹ˆ๋‹ค.

6. AI์™€ ํ•จ๊ป˜ํ•˜๋Š” ์ฆ๊ฑฐ์šด ์ผ์ƒ

์•ผ์‹œ๋Š” AI๋ฅผ ์ฃผ๋กœ ์—…๋ฌด ๋„๊ตฌ๋กœ ํ™œ์šฉํ•˜์ง€๋งŒ, ๊ฐœ์ธ์ ์ธ ์ฆ๊ฑฐ์›€์„ ์œ„ํ•œ ์šฉ๋„๋กœ๋„ ์‚ฌ์šฉํ•ฉ๋‹ˆ๋‹ค. ๊ฐ€์žฅ ์ข‹์•„ํ•˜๋Š” ๊ฒƒ์€ โ€˜๋ธŒ๋ ˆ์ธ์Šคํ† ๋ฐ(brainstorming)โ€˜๊ณผ โ€˜์—ฐ๊ตฌโ€™์ž…๋‹ˆ๋‹ค. ๊ทธ๋Š” ์นœ๊ตฌ๋“ค๊ณผ ๋งค๋…„ ๊ฒจ์šธ, ์—ฌ๋ฆ„, ์‹ฌ์ง€์–ด ๋ด„, ๊ฐ€์„ ์˜ฌ๋ฆผํ”ฝ๊นŒ์ง€ ๊ฐœ์ตœํ•˜๋Š”๋ฐ, ์ด ํ–‰์‚ฌ์—๋Š” ๋‹ค์–‘ํ•œ ๊ฒŒ์ž„๊ณผ ํ™œ๋™์ด ํฌํ•จ๋ฉ๋‹ˆ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด, โ€˜๊ฒ€, ๋ฌผ๊ณ ๊ธฐ, ์ˆ˜ํ”„ ์ค‘ ๋ฌด์—‡์ผ๊นŒ?โ€™ ๊ฐ™์€ ๊ธฐ๋ฐœํ•œ ํ€ด์ฆˆ ๊ฒŒ์ž„์ด๋‚˜, ๋Œ€ํ•™ ์‹œ์ ˆ ํ–ฅ์ˆ˜๋ฅผ ๋ถˆ๋Ÿฌ์ผ์œผํ‚ค๋Š” ์Œ์ฃผ ๊ฒŒ์ž„, ๋Œ€ํ™”ํ˜• ๊ฒŒ์ž„ ๋“ฑ์ž…๋‹ˆ๋‹ค.

๊ทธ๋Š” ์ž๋น„์Šค(Jarvis)์™€ ํ•จ๊ป˜ ๊ณผ๊ฑฐ์— ํ–ˆ๋˜ ํ™œ๋™๋“ค์„ ๋˜์งš์–ด๋ณด๊ณ  ์ƒˆ๋กœ์šด ํ™œ๋™ ์•„์ด๋””์–ด๋ฅผ ๊ตฌ์ƒํ•˜๋ฉฐ, ๊ฐ ๊ฒŒ์ž„์˜ ํ…Œ๋งˆ๋ฅผ ์„ค์ •ํ•˜๋Š” ๋ฐ ๋„์›€์„ ๋ฐ›์Šต๋‹ˆ๋‹ค. AI๊ฐ€ ์™„๋ฒฝํ•œ ๋‹ต์„ ์ฃผ์ง€๋Š” ์•Š์ง€๋งŒ, ๊ทธ์˜ ์‚ฌ๊ณ  ๊ณผ์ •์„ ์ž๊ทนํ•˜์—ฌ ์ตœ์ข…์ ์œผ๋กœ ๋” ์žฌ๋ฏธ์žˆ๋Š” ์•„์ด๋””์–ด๋ฅผ ์™„์„ฑํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋•์Šต๋‹ˆ๋‹ค. ๋˜ํ•œ, 20๋ช…์— ๋‹ฌํ•˜๋Š” ์ฐธ๊ฐ€์ž๋“ค์˜ ํŒ€์„ ๊ตฌ์„ฑํ•˜๊ณ , ๋ชจ๋“  ์ฐธ๊ฐ€์ž๊ฐ€ ์„œ๋กœ ๊ต๋ฅ˜ํ•  ์ˆ˜ ์žˆ๋„๋ก ๊ฒŒ์ž„๋ณ„๋กœ ํŽ˜์–ด๋ฅผ ๋กœํ…Œ์ด์…˜ํ•˜๋Š” ๋ณต์žกํ•œ ์šด์˜(ops) ๊ณ„ํš์„ ์„ธ์šฐ๋Š” ๋ฐ๋„ AI๋ฅผ ํ™œ์šฉํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Š” ์ตœ๊ทผ ํด๋กœ๋“œ(Claude)๋ฅผ ํ†ตํ•ด 10๊ฐœ์˜ ์ƒˆ๋กœ์šด ๋ณด๋“œ๊ฒŒ์ž„์„ ๋ฐœ๊ฒฌํ•˜๊ณ  ๊ตฌ๋งคํ–ˆ์„ ์ •๋„๋กœ, AI๋ฅผ ํ†ตํ•ด ์—ฌ๊ฐ€ ์ƒํ™œ๋„ ํ’์„ฑํ•˜๊ฒŒ ๋งŒ๋“ค๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.

๊ฒฐ๋ก : AI๊ฐ€ ์—ฌ๋Š” ๊ฐœ์ธํ™”๋œ ์—…๋ฌด์™€ ์‚ถ์˜ ์ƒˆ๋กœ์šด ์ง€ํ‰

์•ผ์‹œ ํ…Œํฌ๋ฆฌ์›”์˜ ์‚ฌ๋ก€๋Š” AI๊ฐ€ ๋‹จ์ˆœํžˆ ๋ณต์žกํ•œ ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๋Š” ๋„๊ตฌ๋ฅผ ๋„˜์–ด, ์šฐ๋ฆฌ์˜ ๋””์ง€ํ„ธ ํ™˜๊ฒฝ๊ณผ ์ƒํ˜ธ์ž‘์šฉํ•˜๋Š” ๋ฐฉ์‹์„ ๊ทผ๋ณธ์ ์œผ๋กœ ๋ณ€ํ™”์‹œํ‚ค๊ณ  ์žˆ์Œ์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค. ์Šฌ๋ž™ ์•Œ๋ฆผ ๊ณผ๋ถ€ํ•˜์™€ ๊ฐ™์€ ๋ณดํŽธ์ ์ธ ๋ฌธ์ œ์— ๋Œ€ํ•ด, AI๋Š” ๊ฐœ์ธ์˜ ๊ณ ์œ ํ•œ ๋‹ˆ์ฆˆ์™€ ์›Œํฌํ”Œ๋กœ์šฐ์— ์™„๋ฒฝํ•˜๊ฒŒ ์ตœ์ ํ™”๋œ ๋งž์ถคํ˜• ์†”๋ฃจ์…˜์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.

์ด๋Š” โ€˜์‚ฌ์Šค(SaaS)์˜ ์ข…๋งโ€™์ด ์•„๋‹ˆ๋ผ โ€˜๊ฐœ์ธํ™”๋œ ์‚ฌ์Šค ๊ฒฝํ—˜์˜ ์‹œ์ž‘โ€™์„ ์˜๋ฏธํ•˜๋ฉฐ, ์ €๋ ดํ•œ ๋น„์šฉ์œผ๋กœ ํ‹ˆ์ƒˆ์‹œ์žฅ์„ ๊ณต๋žตํ•˜๋Š” ์ˆ˜๋งŽ์€ ๋งˆ์ดํฌ๋กœ ์‚ฌ์Šค๋“ค์˜ ๋“ฑ์žฅ์„ ์˜ˆ๊ณ ํ•ฉ๋‹ˆ๋‹ค. ๋˜ํ•œ, โ€˜์•ˆํ‹ฐ-ํˆฌ๋‘ ๋ฆฌ์ŠคํŠธโ€™์™€ ๊ฐ™์ด ๋ฐ˜๋ณต์ ์ด๊ณ  ์ง€๋ฃจํ•œ ์—…๋ฌด๋ฅผ AI์—๊ฒŒ ๋งก๊น€์œผ๋กœ์จ, ์šฐ๋ฆฌ๋Š” ๋” ์ฐฝ์˜์ ์ด๊ณ  ๊ฐ€์น˜ ์žˆ๋Š” ์ผ์— ์ง‘์ค‘ํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋ฉ๋‹ˆ๋‹ค.

๊ฐœ์ธ ์ƒ์‚ฐ์„ฑ ํ–ฅ์ƒ์„ ๋„˜์–ด ํŒ€ ํ˜‘์—… ๋ฐฉ์‹์˜ ํ˜์‹ , ๊ทธ๋ฆฌ๊ณ  ์—ฌ๊ฐ€ ์ƒํ™œ์˜ ํ’์š”๋กœ์›€๊นŒ์ง€, AI๋Š” ์šฐ๋ฆฌ๊ฐ€ ๋” ์ ์€ ๋…ธ๋ ฅ์œผ๋กœ ๋” ๋งŽ์€ ๊ฒƒ์„ ์ด๋ฃฐ ์ˆ˜ ์žˆ๋„๋ก ๋•๋Š” ๊ฐ•๋ ฅํ•œ ์กฐ๋ ฅ์ž๊ฐ€ ๋˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ์ด์ œ AI๋Š” ๋” ์ด์ƒ ๋จผ ๋ฏธ๋ž˜์˜ ๊ธฐ์ˆ ์ด ์•„๋‹ˆ๋ผ, ์šฐ๋ฆฌ์˜ ์ผ๊ณผ ์‚ถ์„ ๋”์šฑ ์Šค๋งˆํŠธํ•˜๊ณ  ์ฆ๊ฒ๊ฒŒ ๋งŒ๋“œ๋Š” ํ˜„์žฌ์ง„ํ–‰ํ˜•์˜ ํ˜์‹ ์ž…๋‹ˆ๋‹ค.


โ€œThe Era of AI Agents | Aaron Levie on The a16z Showโ€ โ€” a16z ๊ธฐ๋ฐ˜ ๊ธฐ์‚ฌ ์›๋ณธ ์˜์ƒ ๋ณด๊ธฐ

AI ์—์ด์ „ํŠธ ์‹œ๋Œ€: ์‹ค๋ฆฌ์ฝ˜๋ฐธ๋ฆฌ์˜ ๋‚™๊ด€๋ก , ํ˜„์‹ค์˜ ๋ณต์žก์„ฑ์„ ๋งŒ๋‚˜๋‹ค

์„œ๋ฌธ: AI ์—์ด์ „ํŠธ, ๊ธฐ๋Œ€์™€ ํ˜„์‹ค ์‚ฌ์ด์˜ ์ค„๋‹ค๋ฆฌ๊ธฐ

์ธ๊ณต์ง€๋Šฅ(AI) ์—์ด์ „ํŠธ์˜ ๋“ฑ์žฅ์€ ๋‹จ์ˆœํ•œ ๊ธฐ์ˆ  ๋ฐœ์ „์„ ๋„˜์–ด ์†Œํ”„ํŠธ์›จ์–ด์™€ ๊ธฐ์—… ์šด์˜์˜ ๊ทผ๋ณธ์ ์ธ ๋ณ€ํ™”๋ฅผ ์˜ˆ๊ณ ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. a16z ์‡ผ์— ์ถœ์—ฐํ•œ Box์˜ CEO ์• ๋Ÿฐ ๋ ˆ๋น„(Aaron Levie)์™€ ์ง„ํ–‰์ž์˜ ๋Œ€ํ™”๋Š” ์ด๋Ÿฌํ•œ ๋ณ€ํ™”์˜ ์ž ์žฌ๋ ฅ๊ณผ ํ•จ๊ป˜, ์‹ค๋ฆฌ์ฝ˜๋ฐธ๋ฆฌ์˜ ๋‚™๊ด€๋ก ์ด ๊ฐ„๊ณผํ•˜๊ธฐ ์‰ฌ์šด ํ˜„์‹ค์ ์ธ ๋„์ „ ๊ณผ์ œ๋“ค์„ ์‹ฌ๋„ ์žˆ๊ฒŒ ์กฐ๋ช…ํ•ฉ๋‹ˆ๋‹ค. AI ์—์ด์ „ํŠธ๊ฐ€ ์†Œํ”„ํŠธ์›จ์–ด์˜ ์ƒˆ๋กœ์šด ์ฃผ๋ฅ˜ ์‚ฌ์šฉ์ž๊ฐ€ ๋  ๊ฒƒ์ด๋ผ๋Š” ๋น„์ „๋ถ€ํ„ฐ, ๊ธฐ์—… ํ™˜๊ฒฝ์—์„œ์˜ ๋ณด์•ˆ ๋ฐ ํ†ตํ•ฉ ๋ฌธ์ œ, ๊ทธ๋ฆฌ๊ณ  ๋ฐ์ดํ„ฐ ์ ‘๊ทผ ๋ฐฉ์‹์˜ ๋ณ€ํ™”์— ์ด๋ฅด๊ธฐ๊นŒ์ง€, ์ด๋ฒˆ ๋Œ€๋‹ด์€ AI ์—์ด์ „ํŠธ ์‹œ๋Œ€๊ฐ€ ์šฐ๋ฆฌ์—๊ฒŒ ๋˜์ง€๋Š” ์งˆ๋ฌธ๋“ค์„ ๋ช…ํ™•ํžˆ ์ œ์‹œํ•ฉ๋‹ˆ๋‹ค.

1. AI ์—์ด์ „ํŠธ: ์†Œํ”„ํŠธ์›จ์–ด์˜ ์ƒˆ๋กœ์šด ์ฃผ๋ฅ˜ ์‚ฌ์šฉ์ž

์• ๋Ÿฐ ๋ ˆ๋น„๋Š” โ€œAI ์—ญ๋Ÿ‰์˜ ํ™•์‚ฐ์€ ์‹ค๋ฆฌ์ฝ˜๋ฐธ๋ฆฌ ์‚ฌ๋žŒ๋“ค์ด ์ƒ๊ฐํ•˜๋Š” ๊ฒƒ๋ณด๋‹ค ๋” ์˜ค๋ž˜ ๊ฑธ๋ฆด ๊ฒƒโ€์ด๋ผ๊ณ  ๋‹จ์–ธํ•˜๋ฉฐ ๋Œ€ํ™”๋ฅผ ์‹œ์ž‘ํ•ฉ๋‹ˆ๋‹ค. ๊ทธ ์ด์œ ๋Š” SAP์™€ ๊ฐ™์€ ๋ณต์žกํ•œ ์‹œ์Šคํ…œ์„ ๋‹จ์ˆœํžˆ โ€˜๊ฐ์œผ๋กœ ์ฝ”๋”ฉ(vibe code)โ€˜ํ•˜๋Š” ๊ฒƒ์ด ๋ถˆ๊ฐ€๋Šฅํ•˜๋ฉฐ, ์—„์ฒญ๋‚œ ์–‘์˜ ๋„๋ฉ”์ธ ์ง€์‹(domain knowledge)์ด ์ž˜ ์ •๋ˆ๋œ ๋ฐ์ดํ„ฐ ๊ณ„์ธต์—๋งŒ ์กด์žฌํ•˜๋Š” ๊ฒƒ์ด ์•„๋‹ˆ๊ธฐ ๋•Œ๋ฌธ์ด๋ผ๊ณ  ์„ค๋ช…ํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ๊ธฐํšŒ์˜ ๊ทœ๋ชจ๋Š” ์‚ฌ๋žŒ๋“ค์ด ์ƒ๊ฐํ•˜๋Š” ๊ฒƒ๋ณด๋‹ค ํ›จ์”ฌ ํฌ๋‹ค๊ณ  ๊ฐ•์กฐํ•˜๋ฉฐ, โ€œ์‚ฌ๋žŒ๋ณด๋‹ค 100๋ฐฐ ๋˜๋Š” 1000๋ฐฐ ๋” ๋งŽ์€ ์—์ด์ „ํŠธ๊ฐ€ ์žˆ๋‹ค๋ฉด, ์†Œํ”„ํŠธ์›จ์–ด๋Š” ์—์ด์ „ํŠธ๋ฅผ ์œ„ํ•ด ๋งŒ๋“ค์–ด์ ธ์•ผ ํ•œ๋‹คโ€๋Š” ํ•ต์‹ฌ ์ฃผ์žฅ์„ ํŽผ์นฉ๋‹ˆ๋‹ค.

์—์ด์ „ํŠธ๊ฐ€ ์‹œ์Šคํ…œ๊ณผ ์ƒํ˜ธ์ž‘์šฉํ•˜๋Š” ๋ฐฉ์‹์€ API(Application Programming Interface), CLI(Command Line Interface) ๋“ฑ ๋‹ค์–‘ํ•œ ํ˜•ํƒœ๊ฐ€ ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค. ํŠนํžˆ ์ฝ”๋”ฉ ์—์ด์ „ํŠธ(coding agent)๊ฐ€ SaaS(Software as a Service) ๋„๊ตฌ์™€ ์ง€์‹ ์—…๋ฌด ์›Œํฌํ”Œ๋กœ์— ์ ‘๊ทผํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋˜๋ฉด, ์ด๋Š” ๋งˆ์น˜ ์ดˆ๋Šฅ๋ ฅ(superpower)๊ณผ ๊ฐ™์€ ์—ญํ• ์„ ํ•˜๊ฒŒ ๋ฉ๋‹ˆ๋‹ค. ์—์ด์ „ํŠธ๋Š” ๋‹จ์ˆœํžˆ ๋ฐ์ดํ„ฐ๋ฅผ ์ฝ๊ณ  ์ •๋ณด๋ฅผ ์ดํ•ดํ•˜๋Š” ๊ฒƒ์„ ๋„˜์–ด, ์ฝ”๋“œ๋ฅผ ์ƒ์„ฑํ•˜๊ฑฐ๋‚˜ API๋ฅผ ํ™œ์šฉํ•˜์—ฌ ๋ชฉํ‘œํ•˜๋Š” ์ž‘์—…์„ ๋‹ฌ์„ฑํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋ฉ๋‹ˆ๋‹ค. ์ด๋Š” ํด๋กœ๋“œ ์ฝ”๋“œ(Claude Code), OpenAI์˜ ํ”„๋กœ์ ํŠธ, ํผํ”Œ๋ ‰์‹œํ‹ฐ(Perplexity) ์ปดํ“จํŠธ ๋“ฑ์—์„œ ์ด๋ฏธ ๋‚˜ํƒ€๋‚˜๊ณ  ์žˆ๋Š” ํ˜„์ƒ์ด๋ฉฐ, ์ด๋Ÿฌํ•œ ํŒจ๋Ÿฌ๋‹ค์ž„์€ ๋น ๋ฅด๊ฒŒ ํ™•์‚ฐ๋  ๊ฒƒ์ด๋ผ๊ณ  ์ „๋งํ•ฉ๋‹ˆ๋‹ค.

2. โ€˜์‹œ์Šคํ…œ ์‚ฌ๊ณ โ€™์˜ ์ค‘์š”์„ฑ๊ณผ ์ง๋ฌด์˜ ์ง„ํ™”

ํ•˜์ง€๋งŒ ์—์ด์ „ํŠธ๋ฅผ ํšจ๊ณผ์ ์œผ๋กœ ํ™œ์šฉํ•˜๋Š” ๊ฒƒ์€ ๋ชจ๋‘์—๊ฒŒ ์‰ฌ์šด ์ผ์ด ์•„๋‹™๋‹ˆ๋‹ค. ์ง„ํ–‰์ž๋Š” โ€œ์•Œ๊ณ ๋ฆฌ์ฆ˜์  ์‚ฌ๊ณ (algorithmic thinking)๋Š” ๋Œ€๋ถ€๋ถ„์˜ ์ง์žฅ์ธ์—๊ฒŒ ๋งค์šฐ ์–ด๋ ต๋‹คโ€๊ณ  ์ง€์ ํ•˜๋ฉฐ, ๋Œ€๋ถ€๋ถ„์˜ ์‚ฌ๋žŒ์ด ์ž์‹ ์˜ ์—…๋ฌด์— ๋Œ€ํ•œ ํ”Œ๋กœ์šฐ์ฐจํŠธ(flowchart)๋ฅผ ์ง์ ‘ ๊ทธ๋ฆฌ๋Š” ๊ฒƒ์กฐ์ฐจ ์–ด๋ ค์›Œํ•œ๋‹ค๊ณ  ๋งํ•ฉ๋‹ˆ๋‹ค. ์ด๋Š” ํŠน์ • ์กฐ์ง ๋‚ด์—์„œ ์ „์ฒด ํ”„๋กœ์„ธ์Šค๋ฅผ ์ •ํ™•ํžˆ ์ดํ•ดํ•˜๊ณ  ๋ฌธ์„œํ™”ํ•  ์ˆ˜ ์žˆ๋Š” ์‚ฌ๋žŒ์€ ๊ทนํžˆ ์ผ๋ถ€์— ๋ถˆ๊ณผํ•˜๋‹ค๋Š” ํ˜„์‹ค์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค.

์ด๋Ÿฌํ•œ ๋งฅ๋ฝ์—์„œ ์—์ด์ „ํŠธ์—๊ฒŒ โ€˜๋ฌด์—‡์„ ํ•ด์•ผ ํ• ์ง€โ€™๋ฅผ ์„ค๋ช…ํ•˜๋Š” ๋Šฅ๋ ฅ์€ ๋งค์šฐ ์ œํ•œ์ ์ผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ๋ ˆ๋น„๋Š” ์ด๋Ÿฌํ•œ ๋ณ€ํ™”๊ฐ€ ์ƒˆ๋กœ์šด ์ถ”์ƒํ™” ๊ณ„์ธต(abstraction layer)์„ ๋ฐœ์ „์‹œํ‚ค๋Š” ๊ณผ์ •์ด๋ผ๊ณ  ์„ค๋ช…ํ•ฉ๋‹ˆ๋‹ค. ๊ณผ๊ฑฐ ์Šคํ”„๋ ˆ๋“œ์‹œํŠธ๊ฐ€ ์ฒ˜์Œ ๋„์ž…๋˜์—ˆ์„ ๋•Œ, MBA๋ฅผ ์กธ์—…ํ•œ ๋งค๋‹ˆ์ €๊ฐ€ ์Šคํ”„๋ ˆ๋“œ์‹œํŠธ๋ฅผ ๋‹ค๋ฃฐ ์ค„ ๋ชฐ๋ผ ์ธํ„ด๋“ค์„ ๊ณ ์šฉํ–ˆ๋˜ ์‚ฌ๋ก€๋ฅผ ํ†ตํ•ด, ๊ฒฐ๊ตญ ๋งค๋‹ˆ์ €๋“ค์ด ์ง์ ‘ ์Šคํ”„๋ ˆ๋“œ์‹œํŠธ๋ฅผ ๋‹ค๋ฃจ๊ฒŒ ๋˜๋ฉด์„œ ์ง๋ฌด๊ฐ€ ํ•œ ๋‹จ๊ณ„ ์ƒ์Šน(job moves up a rung)ํ–ˆ๋˜ ๊ฒƒ์ฒ˜๋Ÿผ, AI ์—์ด์ „ํŠธ ์‹œ๋Œ€์—๋„ ์œ ์‚ฌํ•œ ๋ณ€ํ™”๊ฐ€ ์ผ์–ด๋‚  ๊ฒƒ์ด๋ผ๋Š” ์ „๋ง์ž…๋‹ˆ๋‹ค.

์‹ค์ œ๋กœ ์•คํŠธ๋กœํ”ฝ(Anthropic)์˜ ์„ฑ์žฅ ๋งˆ์ผ€ํ„ฐ ์‚ฌ๋ก€๋Š” ํ•œ ์‚ฌ๋žŒ์ด ํด๋กœ๋“œ ์ฝ”๋“œ๋ฅผ ํ™œ์šฉํ•˜์—ฌ 5~10๋ช…์˜ ์—…๋ฌด๋ฅผ ์ž๋™ํ™”ํ•˜๋Š” ๋ชจ์Šต์„ ๋ณด์—ฌ์ฃผ์—ˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์‹œ์Šคํ…œ ์‚ฌ์ƒ๊ฐ€(systems thinker)์˜ ์ค‘์š”์„ฑ์„ ๊ฐ•์กฐํ•˜๋ฉฐ, ๋ฏธ๋ž˜์˜ ์ง์—…์€ ๋ฌดํ•œํ•œ ์—”์ง€๋‹ˆ์–ด ํ’€(pool of engineers)์„ ํ™œ์šฉํ•˜์—ฌ ์—…๋ฌด๋ฅผ ์ž๋™ํ™”ํ•  ์ˆ˜ ์žˆ๋Š” โ€˜์‹œ์Šคํ…œ์  ์‚ฌ๊ณ โ€™๋ฅผ ์š”๊ตฌํ•˜๊ฒŒ ๋  ๊ฒƒ์ด๋ผ๊ณ  ๋งํ•ฉ๋‹ˆ๋‹ค. ์—์ด์ „ํŠธ๊ฐ€ ์ด๋Ÿฌํ•œ ๋ฐฉํ–ฅ์œผ๋กœ ์ธ๊ฐ„์„ ์•ˆ๋‚ดํ•  ์ˆ˜ ์žˆ๋„๋ก ์ง„ํ™”ํ•  ๊ฒƒ์ด๋ผ๋Š” ๊ธฐ๋Œ€๋„ ๋ง๋ถ™์˜€์Šต๋‹ˆ๋‹ค.

3. ์—์ด์ „ํŠธ์™€ ์†Œํ”„ํŠธ์›จ์–ด ์ƒํ˜ธ์ž‘์šฉ์˜ ์ง„ํ™”: ์ฝ”๋“œ์—์„œ ์ปดํ“จํ„ฐ ์‚ฌ์šฉ์œผ๋กœ

์—์ด์ „ํŠธ์™€ ์†Œํ”„ํŠธ์›จ์–ด์˜ ์ƒํ˜ธ์ž‘์šฉ ๋ฐฉ์‹์— ๋Œ€ํ•œ ๋…ผ์˜๋„ ํฅ๋ฏธ๋กœ์šด ์ง€์ ์ž…๋‹ˆ๋‹ค. ๋ ˆ๋น„๋Š” ์ดˆ๊ธฐ์—๋Š” AI๋ฅผ SaaS ์†Œํ”„ํŠธ์›จ์–ด์— ํ†ตํ•ฉํ•˜๋Š” ๋ฐฉ์‹(AI-enabled SaaS)์— ์ง‘์ค‘ํ–ˆ์ง€๋งŒ, ์ด์ œ๋Š” SaaS ์†Œํ”„ํŠธ์›จ์–ด๊ฐ€ ์—์ด์ „ํŠธ์˜ โ€˜์ปดํ“จํ„ฐโ€™ ์—ญํ• ์„ ํ•˜๋Š” ๋ฐฉํ–ฅ์œผ๋กœ ์ง„ํ™”ํ•˜๊ณ  ์žˆ๋‹ค๊ณ  ๋งํ•ฉ๋‹ˆ๋‹ค. ์ฆ‰, โ€˜์ฝ”๋“œ ์ƒ์„ฑโ€™์—์„œ โ€˜ํ„ฐ๋ฏธ๋„(terminal) ์‚ฌ์šฉโ€™์„ ๊ฑฐ์ณ โ€˜์ปดํ“จํ„ฐ ์‚ฌ์šฉ(computer use)โ€˜์˜ ์‹œ๋Œ€๋กœ ์ ‘์–ด๋“ค๊ณ  ์žˆ๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค. ์—์ด์ „ํŠธ๊ฐ€ ๋งˆ์น˜ ์ธ๊ฐ„์ฒ˜๋Ÿผ ์ปดํ“จํ„ฐ๋ฅผ ํ™œ์šฉํ•˜๋Š” ๋ฐฉ์‹์ด ๋”์šฑ ์ค‘์š”ํ•ด์งˆ ๊ฒƒ์ด๋ผ๋Š” ์˜๋ฏธ์ž…๋‹ˆ๋‹ค.

๊ทธ๋Ÿฌ๋‚˜ ์ง„ํ–‰์ž๋Š” ์ฝ”๋“œ ์ƒ์„ฑ, API ์‚ฌ์šฉ, ์ปดํ“จํ„ฐ ์‚ฌ์šฉ์„ ํ•˜๋‚˜์˜ ๋ฒ”์ฃผ๋กœ ๋ฌถ๋Š” ๊ฒƒ์— ์ด์˜๋ฅผ ์ œ๊ธฐํ•˜๋ฉฐ, ์ด๋“ค์ด ๋งค์šฐ ๋‹ค๋ฅธ ๊ฐœ๋…์ด๋ผ๊ณ  ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค. ์ด์— ๋ ˆ๋น„๋Š” Box์˜ ๋‚ด๋ถ€ ์—์ด์ „ํŠธ ์‚ฌ๋ก€๋ฅผ ๋“ค์–ด ๋ฐ˜๋ฐ•ํ•ฉ๋‹ˆ๋‹ค. ์ด ์—์ด์ „ํŠธ๋Š” ๊ธฐ์กด ์Šคํ‚ฌ, Box์˜ ๋„๊ตฌ, ๋˜๋Š” ์ฝ”๋“œ๋ฅผ ์ง์ ‘ ์ž‘์„ฑํ•˜๋Š” ์„ธ ๊ฐ€์ง€ ์˜ต์…˜ ์ค‘ ๊ฐ€์žฅ ์ ์ ˆํ•œ ๊ฒƒ์„ ์Šค์Šค๋กœ ํŒ๋‹จํ•˜์—ฌ ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•ฉ๋‹ˆ๋‹ค. ๋ฌธ์„œ์— ๋Œ€ํ•œ ๋ชจ๋“  ๊ฐ€๋Šฅํ•œ ์ž‘์—…์„ ๋ฏธ๋ฆฌ ๊ณ„ํšํ•  ์ˆ˜ ์—†์œผ๋ฏ€๋กœ, ํ•„์š”์— ๋”ฐ๋ผ ์ฝ”๋“œ๋ฅผ ์ฆ‰์„์—์„œ ์ž‘์„ฑํ•˜๋Š” ์—์ด์ „ํŠธ์˜ ๋Šฅ๋ ฅ์€ ๋งค์šฐ ์œ ์šฉํ•˜๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค.

์ด๋Š” ์ธ๊ฐ„์ด ์—ฌ๋Ÿฌ ์•ฑ์„ ๋ฐฐ์šฐ๋Š” ๋ฐ ๋“œ๋Š” ์ •์‹ ์  ์ œ์•ฝ์ด ์—์ด์ „ํŠธ์—๊ฒŒ๋Š” ์—†๊ธฐ ๋•Œ๋ฌธ์—, ์—์ด์ „ํŠธ๋Š” ๋‹ค์–‘ํ•œ ๋„๊ตฌ์™€ API๋ฅผ ํ™œ์šฉํ•˜๊ณ  ์ฝ”๋“œ๋ฅผ ์ƒ์„ฑํ•˜๋Š” ๋ฐ ์•„๋ฌด๋Ÿฐ ์ œ์•ฝ์ด ์—†๋‹ค๋Š” ์ ์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค. ๊ฒฐ๊ตญ ์—์ด์ „ํŠธ์˜ ๋“ฑ์žฅ์œผ๋กœ ์ธ๊ฐ„ ์‚ฌ์šฉ์ž์˜ ๋ณ‘๋ชฉ ํ˜„์ƒ์ด ํ•ด์†Œ๋˜๊ณ , ์ง€๋‚œ 25๋…„๊ฐ„ ์†Œํ”„ํŠธ์›จ์–ด ๊ธฐ๋Šฅ์˜ ์ž ์žฌ๋ ฅ์„ ์˜จ์ „ํžˆ ํ™œ์šฉํ•  ์ˆ˜ ์žˆ๊ฒŒ ๋  ๊ฒƒ์ด๋ผ๋Š” ์ „๋ง์ž…๋‹ˆ๋‹ค.

4. ๊ธฐ์—… ํ™˜๊ฒฝ์—์„œ์˜ ๋„์ „: ๋ณด์•ˆ, ๊ฑฐ๋ฒ„๋„Œ์Šค, ๊ทธ๋ฆฌ๊ณ  ํ†ต์ œ

๊ฐœ์ธ ์ƒ์‚ฐ์„ฑ ์ธก๋ฉด์—์„œ AI ์—์ด์ „ํŠธ์˜ ์ž ์žฌ๋ ฅ์€ ๋ถ„๋ช…ํ•˜์ง€๋งŒ, ๊ธฐ์—… ํ™˜๊ฒฝ์—์„œ์˜ ๋„์ž…์€ ์ˆ˜๋งŽ์€ ๋ณต์žกํ•œ ๋ฌธ์ œ๋ฅผ ์•ผ๊ธฐํ•ฉ๋‹ˆ๋‹ค. CFO์™€ CIO๋“ค์€ ์—์ด์ „ํŠธ๊ฐ€ ์‹œ์Šคํ…œ ํ†ตํ•ฉ(integration)์„ ๋” ์‰ฝ๊ฒŒ ๋งŒ๋“ค ๊ฒƒ์ด๋ผ๋Š” ์ฃผ์žฅ์— ๋Œ€ํ•ด โ€œ์‹œ์Šคํ…œ ๊ธฐ๋ก์„ ๋ง๊ฐ€๋œจ๋ฆด ์œ„ํ—˜์ด ์žˆ๋‹คโ€๋ฉฐ ๊ฐ•๋ ฅํ•œ ์šฐ๋ ค๋ฅผ ํ‘œํ•ฉ๋‹ˆ๋‹ค. ์ƒˆ๋กœ์šด ํ†ตํ•ฉ์„ ๋ฌด๋ถ„๋ณ„ํ•˜๊ฒŒ ์ƒ์„ฑํ•˜๋Š” ๊ฒƒ์€ ๊ธฐ์—… ์‹œ์Šคํ…œ์— ์น˜๋ช…์ ์ธ ์˜ํ–ฅ์„ ๋ฏธ ์ค„ ์ˆ˜ ์žˆ๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค.

Box CLI ์‚ฌ๋ก€์—์„œ๋„ ์—์ด์ „ํŠธ๊ฐ€ ๋ฌดํ•œ ๋ฃจํ”„์— ๋น ์ ธ ๋””๋ ‰ํ„ฐ๋ฆฌ๋ฅผ ๋ฌดํ•œ์ • ์ƒ์„ฑํ•˜๋Š” ๋ฌธ์ œ๊ฐ€ ๋ฐœ์ƒํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์—์ด์ „ํŠธ๊ฐ€ ์˜๋„์น˜ ์•Š๊ฒŒ ์‹œ์Šคํ…œ์— ๊ณผ๋ถ€ํ•˜๋ฅผ ์ฃผ๊ฑฐ๋‚˜ ๋ฐ์ดํ„ฐ๋ฅผ ์†์ƒ์‹œํ‚ฌ ์ˆ˜ ์žˆ์Œ์„ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค. ๊ฐœ์ธ ์—์ด์ „ํŠธ์—๊ฒŒ๋Š” ์ž์ฒด ์ „ํ™”๋ฒˆํ˜ธ, ์‹ ์šฉ์นด๋“œ, Gmail ๊ณ„์ •์„ ๋ถ€์—ฌํ•˜์—ฌ ๋…๋ฆฝ์ ์ธ ์ฃผ์ฒด์ฒ˜๋Ÿผ ๋‹ค๋ฃฐ ์ˆ˜ ์žˆ์ง€๋งŒ, ๊ธฐ์—… ํ™˜๊ฒฝ์—์„œ๋Š” ์—์ด์ „ํŠธ๋ฅผ โ€˜๋ณ„๊ฐœ์˜ ์ธ๊ฐ„โ€™์ฒ˜๋Ÿผ ์™„๋ฒฝํ•˜๊ฒŒ ๋Œ€์šฐํ•˜๊ธฐ ์–ด๋ ต์Šต๋‹ˆ๋‹ค.

๋ ˆ๋น„๋Š” ๊ธฐ์—…์ด ์—์ด์ „ํŠธ์— ๋Œ€ํ•œ ์™„์ „ํ•œ ๊ฐ๋…(complete oversight)์„ ์œ ์ง€ํ•ด์•ผ ํ•˜๋ฉฐ, ์—์ด์ „ํŠธ์—๊ฒŒ๋Š” ํ”„๋ผ์ด๋ฒ„์‹œ ๊ถŒ๋ฆฌ๊ฐ€ ์—†๋‹ค๊ณ  ์ฃผ์žฅํ•ฉ๋‹ˆ๋‹ค. ์—์ด์ „ํŠธ๋Š” ์ธ๊ฐ„๋ณด๋‹ค ํ›จ์”ฌ ์‰ฝ๊ฒŒ ์ •๋ณด๋ฅผ ์œ ์ถœํ•˜๊ฑฐ๋‚˜ ์‚ฌํšŒ ๊ณตํ•™์  ๊ณต๊ฒฉ(social engineering attack)์— ์ทจ์•ฝํ•  ์ˆ˜ ์žˆ๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค. ์ด๋Š” ์—์ด์ „ํŠธ๊ฐ€ ๋‹จ์ˆœํ•œ ๋„๊ตฌ๊ฐ€ ์•„๋‹Œ โ€˜์‚ฌ์šฉ์ž์˜ ํ™•์žฅ(extension of you)โ€˜์ด๋ผ๋Š” ๋ณธ์งˆ์ ์ธ ํ•œ๊ณ„์—์„œ ๋น„๋กฏ๋ฉ๋‹ˆ๋‹ค.

์ด๋Ÿฌํ•œ ๋ฌธ์ œ ํ•ด๊ฒฐ์„ ์œ„ํ•ด ๋ ˆ๋น„๋Š” ์˜คํ”ˆ์†Œ์Šค(open source)์˜ ์‚ฌ๋ก€๋ฅผ ์–ธ๊ธ‰ํ•ฉ๋‹ˆ๋‹ค. ์˜คํ”ˆ์†Œ์Šค๊ฐ€ ์ฒ˜์Œ ๋“ฑ์žฅํ–ˆ์„ ๋•Œ ๋ผ์ด์„ ์‹ฑ, ํ’ˆ์งˆ ๋“ฑ์˜ ๋ฌธ์ œ๋กœ ๋งŽ์€ ๋…ผ์˜๊ฐ€ ์žˆ์—ˆ๊ณ , ๊ฒฐ๊ตญ ์—…๊ณ„ ํ‘œ์ค€๊ณผ ๊ทœ๋ฒ”์ด ๋งˆ๋ จ๋˜์—ˆ์Šต๋‹ˆ๋‹ค. AI ์—์ด์ „ํŠธ ์—ญ์‹œ ์ด๋Ÿฌํ•œ ๋…ผ์˜ ๊ณผ์ •์„ ๊ฑฐ์ณ ์ƒˆ๋กœ์šด ํ‘œ์ค€๊ณผ ํ†ต์ œ ๋ฉ”์ปค๋‹ˆ์ฆ˜์ด ํ•„์š”ํ•  ๊ฒƒ์ด๋ผ๋Š” ์˜ˆ์ธก์ž…๋‹ˆ๋‹ค.

5. ๊ธฐ์—…์˜ ๋”œ๋ ˆ๋งˆ: ๋ฐ์ดํ„ฐ ์ ‘๊ทผ๊ณผ ๋น„์ฆˆ๋‹ˆ์Šค ๋ชจ๋ธ์˜ ๋ณ€ํ™”

AI ์—์ด์ „ํŠธ์˜ ๋“ฑ์žฅ์€ ๊ธฐ์กด SaaS ๋ฒค๋”๋“ค์˜ ๋น„์ฆˆ๋‹ˆ์Šค ๋ชจ๋ธ์—๋„ ํฐ ๋ณ€ํ™”๋ฅผ ์š”๊ตฌํ•ฉ๋‹ˆ๋‹ค. ํ˜„์žฌ SaaS ๋ฒค๋”๋“ค์€ ๋น„์ฆˆ๋‹ˆ์Šค ๋ผ์ธ ๋ฐ์ดํ„ฐ(line of business data)๋ฟ๋งŒ ์•„๋‹ˆ๋ผ, ๊ทธ ๋ฐ์ดํ„ฐ์— ๊ธฐ๋ฐ˜ํ•œ ์ง€๋Šฅ(intelligence)๊ณผ ๋„๋ฉ”์ธ ์ „๋ฌธ์„ฑ(domain expertise)์„ ํŒ๋งคํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์—์ด์ „ํŠธ๋Š” ์›์‹œ ๋ฐ์ดํ„ฐ(raw data)์— ๋Œ€ํ•œ ๋ฌด์ œํ•œ ์ ‘๊ทผ๊ณผ ๋ผ์ด์„ ์Šค๋ฅผ ์›ํ•ฉ๋‹ˆ๋‹ค. ์ด๋Š” Workday, SAP, Salesforce์™€ ๊ฐ™์€ ๊ธฐ์—…๋“ค์ด ์˜ค๋žซ๋™์•ˆ API ์ ‘๊ทผ์„ฑ ๋ฌธ์ œ๋กœ ์”จ๋ฆ„ํ•ด ์˜จ ์ง€์ ์ž…๋‹ˆ๋‹ค.

SAP๋ฅผ ์˜ˆ๋กœ ๋“ค๋ฉด, ๊ทธ๋“ค์˜ ๋ฐฉ๋Œ€ํ•œ ๋„๋ฉ”์ธ ์ง€์‹์€ ๋‹จ์ˆœํžˆ ์ž˜ ์ •๋ˆ๋œ ๋ฐ์ดํ„ฐ ๊ณ„์ธต์—๋งŒ ์žˆ๋Š” ๊ฒƒ์ด ์•„๋‹ˆ๋ผ UI, ๋ฏธ๋“ค ํ‹ฐ์–ด(middle tiers), ๊ทธ๋ฆฌ๊ณ  ์‹œ์Šคํ…œ ์‚ฌ์šฉ ๋ฐฉ์‹ ์ „๋ฐ˜์— ๊ฑธ์ณ ํผ์ ธ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Ÿฌํ•œ ๋ณต์žก์„ฑ ๋•Œ๋ฌธ์— SAP์™€ ๊ฐ™์€ ๋ ˆ๊ฑฐ์‹œ ์‹œ์Šคํ…œ์—์„œ์˜ AI ํ™•์‚ฐ์€ ๋”๋”œ ์ˆ˜๋ฐ–์— ์—†์Šต๋‹ˆ๋‹ค.

๊ทธ๋Ÿฌ๋‚˜ ๋ ˆ๋น„๋Š” โ€œ์—์ด์ „ํŠธ๊ฐ€ ์›ํ•˜๋Š” ๊ฒƒ์„ ๋งŒ๋“ค์–ด๋ผ(build something agents want)โ€œ๋Š” ํด ๊ทธ๋ ˆ์ด์—„(Paul Graham)์˜ ์•„์ด๋””์–ด๋ฅผ ์ธ์šฉํ•˜๋ฉฐ, ๊ฒฐ๊ตญ ์—์ด์ „ํŠธ๊ฐ€ ์–ด๋–ค ๋„๊ตฌ๋ฅผ ์‚ฌ์šฉํ• ์ง€ ๊ฒฐ์ •ํ•˜๋Š” ์ฃผ์ฒด๊ฐ€ ๋  ๊ฒƒ์ด๋ผ๊ณ  ๋งํ•ฉ๋‹ˆ๋‹ค. ์ถฉ๋ถ„ํ•œ ๋ฐ˜๋ณต์„ ๊ฑฐ์น˜๋ฉด ์—์ด์ „ํŠธ๋Š” โ€œ๋ ˆ๊ฑฐ์‹œ HR ์‹œ์Šคํ…œ์„ ๊ต์ฒดํ•ด์•ผ ํ•œ๋‹ค. ๊ทธ๋ ‡์ง€ ์•Š์œผ๋ฉด ์ด ์›Œํฌํ”Œ๋กœ๋ฅผ ์ž๋™ํ™”ํ•  ์ˆ˜ ์—†๋‹คโ€๊ณ  ์š”๊ตฌํ•  ์ˆ˜ ์žˆ๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค. ์—์ด์ „ํŠธ๊ฐ€ ์‚ฌ๋žŒ๋ณด๋‹ค 100๋ฐฐ, 1000๋ฐฐ ๋” ๋งŽ์€ ์†Œํ”„ํŠธ์›จ์–ด ๋ณผ๋ฅจ์„ ์ƒ์„ฑํ•œ๋‹ค๋ฉด, ์†Œํ”„ํŠธ์›จ์–ด ์Šคํƒ์€ ๊ฒฐ๊ตญ ์—์ด์ „ํŠธ๋ฅผ ์œ„ํ•ด ๊ตฌ์ถ•๋˜์–ด์•ผ ํ•˜๋ฉฐ, ๊ธฐ์—…์˜ IT ์Šคํƒ ๋˜ํ•œ ์ด๋ฅผ ์ง€์›ํ•˜๋„๋ก ๋ณ€ํ™”ํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค. ์ด๋Š” ๊ณ ํ’ˆ์งˆ API ๊ตฌ์ถ•, ์ƒˆ๋กœ์šด ์ˆ˜์ตํ™” ๋ชจ๋ธ, ์—์ด์ „ํŠธ ์‹ ์› ๋ฐ ์ ‘๊ทผ ํ†ต์ œ(access controls) ๊ด€๋ฆฌ๊ฐ€ ์†Œํ”„ํŠธ์›จ์–ด ๊ธฐ์—…์˜ ํ•ต์‹ฌ ๊ณผ์ œ๊ฐ€ ๋  ๊ฒƒ์ž„์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค.

6. โ€˜๋” ๋‚˜์€ ์‹œ์Šคํ…œโ€™ ๊ตฌ์ถ•์˜ ์‹œ๋Œ€

์• ๋Ÿฐ ๋ ˆ๋น„๋Š” โ€œ์—์ด์ „ํŠธ์—๊ฒŒ โ€˜๋งˆ์ผ€ํŒ…โ€™์„ ํ•ด์•ผ ํ•œ๋‹คโ€๋Š” ์ผ๋ฐ˜์ ์ธ ์ธ์‹์— ๋ฐ˜๋ฐ•ํ•˜๋ฉฐ, ์—์ด์ „ํŠธ๊ฐ€ ๊ถ๊ทน์ ์œผ๋กœ ์ถ”๊ตฌํ•˜๋Š” ๊ฒƒ์€ โ€˜๋” ๋‚˜์€ ์‹œ์Šคํ…œโ€™์ด๋ผ๊ณ  ์ฃผ์žฅํ•ฉ๋‹ˆ๋‹ค. ์ฆ‰, ์—์ด์ „ํŠธ๋Š” ํ™”๋ คํ•œ ์ธํ„ฐํŽ˜์ด์Šค๋‚˜ ๋งˆ์ผ€ํŒ… ๋ฉ”์‹œ์ง€์— ํ˜„ํ˜น๋˜์ง€ ์•Š๊ณ , โ€˜์˜๋ฏธ(semantics)โ€™, โ€˜๋น„์šฉ ๋งค๊ฐœ๋ณ€์ˆ˜(cost parameters)โ€™, โ€˜๋‚ด๊ตฌ์„ฑ(durability)โ€˜๊ณผ ๊ฐ™์€ ๋ณธ์งˆ์ ์ธ ์š”์†Œ๋ฅผ ๊ธฐ๋ฐ˜์œผ๋กœ ๊ธฐ์ˆ ์„ ์„ ํƒํ•  ๊ฒƒ์ด๋ผ๋Š” ์„ค๋ช…์ž…๋‹ˆ๋‹ค.

๊ณผ๊ฑฐ์—๋Š” ๊ฐ€ํŠธ๋„ˆ(Gartner)์™€ ๊ฐ™์€ ์ปจ์„คํŒ… ๊ธฐ๊ด€์ด ์–ด๋–ค ์‹œ์Šคํ…œ์„ ์‚ฌ์šฉํ• ์ง€ ๊ฒฐ์ •ํ•˜๋Š” ๋ฐ ํฐ ์˜ํ–ฅ์„ ๋ฏธ์ณค์ง€๋งŒ, ๋ฏธ๋ž˜์—๋Š” ์—์ด์ „ํŠธ๊ฐ€ ํŠน์ • ์œ ํ˜•์˜ ์ž‘์—…์— ์–ด๋–ค ๋ฐ์ดํ„ฐ๋ฒ ์ด์Šค๊ฐ€ ์ ํ•ฉํ•œ์ง€๋ฅผ ์Šค์Šค๋กœ ํŒ๋‹จํ•˜๊ณ  ์ถ”์ฒœํ•  ๊ฒƒ์ž…๋‹ˆ๋‹ค. ์ด๋Š” ๊ธฐ์ˆ  ์„ ํƒ์˜ ์‹ค๋ ฅ์ฃผ์˜(meritocracy)๋ฅผ ์ด‰์ง„ํ•  ์ˆ˜ ์žˆ์ง€๋งŒ, ๋ ˆ๋น„๋Š” ์‹ค๋ฆฌ์ฝ˜๋ฐธ๋ฆฌ๊ฐ€ ๊ด‘๊ณ ๋‚˜ ์ถ”์ฒœ ๊ตฌ๋งค ๋“ฑ์„ ํ†ตํ•ด ์ด๋Ÿฌํ•œ ์‹ค๋ ฅ์ฃผ์˜๋ฅผ ๋น ๋ฅด๊ฒŒ ํ›ผ์†ํ•  ๊ฐ€๋Šฅ์„ฑ๋„ ๊ฒฝ๊ณ ํ•ฉ๋‹ˆ๋‹ค.

๊ฒฐ๋ก : AI ์—์ด์ „ํŠธ์™€ ํ•จ๊ป˜ ์—ด๋ฆฌ๋Š” ์ƒˆ๋กœ์šด ์‹œ๋Œ€

AI ์—์ด์ „ํŠธ ์‹œ๋Œ€๋Š” ์†Œํ”„ํŠธ์›จ์–ด ๊ฐœ๋ฐœ, ์—…๋ฌด ๋ฐฉ์‹, ๊ทธ๋ฆฌ๊ณ  ๊ธฐ์—… ์šด์˜ ์ „๋ฐ˜์— ๊ฑธ์ณ ํ˜๋ช…์ ์ธ ๋ณ€ํ™”๋ฅผ ๊ฐ€์ ธ์˜ฌ ์ž ์žฌ๋ ฅ์„ ์ง€๋‹ˆ๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์‹ค๋ฆฌ์ฝ˜๋ฐธ๋ฆฌ์˜ ๋‚™๊ด€๋ก ์ด ๊ฐ„๊ณผํ•˜๊ธฐ ์‰ฌ์šด ํ˜„์‹ค์ ์ธ ๋ณต์žก์„ฑ, ์ฆ‰ ๋ณด์•ˆ, ํ†ตํ•ฉ, ๊ฑฐ๋ฒ„๋„Œ์Šค, ๋น„์ฆˆ๋‹ˆ์Šค ๋ชจ๋ธ์˜ ๋ณ€ํ™”์™€ ๊ฐ™์€ ๋ฌธ์ œ๋“ค์€ ์ด ์ „ํ™˜ ๊ณผ์ •์„ ์˜ˆ์ƒ๋ณด๋‹ค ๋”๋””๊ฒŒ ๋งŒ๋“ค ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

์Šคํƒ€ํŠธ์—…์€ ๋ ˆ๊ฑฐ์‹œ ์‹œ์Šคํ…œ์˜ ์ œ์•ฝ ์—†์ด ๋น ๋ฅด๊ฒŒ AI ์—์ด์ „ํŠธ๋ฅผ ๋„์ž…ํ•˜๊ณ  ์ƒˆ๋กœ์šด ๋น„์ฆˆ๋‹ˆ์Šค ๋ชจ๋ธ์„ ์ฐฝ์ถœํ•  ์ˆ˜ ์žˆ๋Š” ๋ฐ˜๋ฉด, ๋Œ€๊ธฐ์—…์€ ๊ธฐ์กด ์‹œ์Šคํ…œ๊ณผ์˜ ํ†ตํ•ฉ, ๋ณด์•ˆ ์œ„ํ—˜ ๊ด€๋ฆฌ, ์ƒˆ๋กœ์šด ๊ทœ๋ฒ”๊ณผ ํ‘œ์ค€ ๋งˆ๋ จ์— ๋” ๋งŽ์€ ์‹œ๊ฐ„๊ณผ ๋…ธ๋ ฅ์„ ๋“ค์—ฌ์•ผ ํ•  ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๊ถ๊ทน์ ์œผ๋กœ๋Š” ์—์ด์ „ํŠธ ์ค‘์‹ฌ์˜ ์†Œํ”„ํŠธ์›จ์–ด ์Šคํƒ์œผ๋กœ์˜ ์ „ํ™˜์ด ๋ถˆ๊ฐ€ํ”ผํ•˜๋ฉฐ, ์ด๋Š” ๊ณ ํ’ˆ์งˆ API์™€ ๊ฐ•๋ ฅํ•œ ์‹œ์Šคํ…œ์„ ๊ตฌ์ถ•ํ•˜๋Š” ๊ฒƒ์ด ๊ธฐ์—…์˜ ์ƒ์กด๊ณผ ์„ฑ๊ณต์„ ๊ฒฐ์ •ํ•˜๋Š” ํ•ต์‹ฌ ์š”์†Œ๊ฐ€ ๋  ๊ฒƒ์ž„์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค. AI ์—์ด์ „ํŠธ์™€ ํ•จ๊ป˜ ์—ด๋ฆฌ๋Š” ์ƒˆ๋กœ์šด ์‹œ๋Œ€๋Š” ์šฐ๋ฆฌ์—๊ฒŒ ๊ธฐ์ˆ ์  ํ˜์‹ ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ, ์‚ฌ๊ณ ๋ฐฉ์‹๊ณผ ๋น„์ฆˆ๋‹ˆ์Šค ๋ชจ๋ธ์˜ ๊ทผ๋ณธ์ ์ธ ์žฌ์ •๋ฆฝ์„ ์š”๊ตฌํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.


โ€œA Cease-Fire in Iranโ€ โ€” New York Times Podcasts ๊ธฐ๋ฐ˜ ๊ธฐ์‚ฌ ์›๋ณธ ์˜์ƒ ๋ณด๊ธฐ

์ผ์ด‰์ฆ‰๋ฐœ ์ด๋ž€ ์œ„๊ธฐ, โ€˜๋ฌธ๋ช… ๋ง์‚ดโ€™ ๊ฒฝ๊ณ  ํ›„ ๊ทน์  ํœด์ „โ€ฆ ๊ทธ๋Ÿฌ๋‚˜ ๋๋‚˜์ง€ ์•Š์€ ๊ฐˆ๋“ฑ์˜ ๊ทธ๋ฆผ์ž

[๋‰ด์š•ํƒ€์ž„์Šค ํŒŸ์บ์ŠคํŠธ โ€˜๋” ๋ฐ์ผ๋ฆฌโ€™ ๋ถ„์„]

2024๋…„ 4์›” 8์ผ ์ˆ˜์š”์ผ, ์ „ ์„ธ๊ณ„๊ฐ€ ์ˆจ์ฃฝ์ด๋ฉฐ ์ง€์ผœ๋ณด๋˜ ์ค‘๋™์˜ ์ผ์ด‰์ฆ‰๋ฐœ ์ƒํ™ฉ์ด ๊ทน์ ์ธ ์ „ํ™˜์ ์„ ๋งž์•˜์Šต๋‹ˆ๋‹ค. ๋ฏธ๊ตญ๊ณผ ์ด๋ž€์ด ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์ด ์ œ์‹œํ•œ ์ตœํ›„ํ†ต์ฒฉ ์‹œํ•œ์„ ๋ถˆ๊ณผ ๋ช‡ ์‹œ๊ฐ„ ์•ž๋‘๊ณ  ์ผ์‹œ์ ์ธ ํœด์ „(ceasefire)์— ํ•ฉ์˜ํ•œ ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๋‰ด์š•ํƒ€์ž„์Šค โ€˜๋” ๋ฐ์ผ๋ฆฌโ€™๋Š” ๋ฐฑ์•…๊ด€ ์ˆ˜์„ ํŠนํŒŒ์› ๋ฐ์ด๋น„๋“œ ์ƒ์–ด(David Sanger)์™€์˜ ์‹ฌ์ธต ์ธํ„ฐ๋ทฐ๋ฅผ ํ†ตํ•ด ์ด ๋ง‰ํŒ ํ•ฉ์˜์— ์ด๋ฅด๊ฒŒ ๋œ ๋ฐฐ๊ฒฝ๊ณผ ๊ทธ ์˜๋ฏธ, ๊ทธ๋ฆฌ๊ณ  ์•ž์œผ๋กœ ํ•ด๊ฒฐํ•ด์•ผ ํ•  ๋‚œ์ œ๋“ค์„ ์กฐ๋ช…ํ–ˆ์Šต๋‹ˆ๋‹ค.

๊ทน์  ํœด์ „: ์ˆจ ๋ง‰ํžˆ๋Š” ๊ธด์žฅ ์† ์ฐพ์•„์˜จ ํ•œ์ˆจ ๋Œ๋ฆด ์‹œ๊ฐ„

์ด๋‚  ์˜คํ›„ 6์‹œ ์งํ›„, ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ ๋ฏธ๊ตญ๊ณผ ์ด๋ž€์ด 2์ฃผ๊ฐ„์˜ ํœด์ „์— ํ•ฉ์˜ํ–ˆ๋‹ค๊ณ  ๋ฐœํ‘œํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด๋ž€ ๊ตญ์˜ ์–ธ๋ก  ์—ญ์‹œ ์ตœ๊ณ  ์ง€๋„์ž์˜ ๋ช…๋ น์— ๋”ฐ๋ผ ๋ชจ๋“  ๊ตฐ์‚ฌ ๋ถ€๋Œ€๊ฐ€ ๋ฐœํฌ๋ฅผ ์ค‘๋‹จํ•  ๊ฒƒ์„ ์ง€์‹œํ•˜๋Š” ์„ฑ๋ช…์„ ๋ฐœํ‘œํ•˜๋ฉฐ ํ•ฉ์˜๋ฅผ ํ™•์ธํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์ด๋ž€๊ณผ์˜ ์ „๋ฉด์ „์œผ๋กœ ์ด์–ด์งˆ ์ˆ˜ ์žˆ์—ˆ๋˜ ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์˜ 8์‹œ(๋™๋ถ€ ํ‘œ์ค€์‹œ) ๋งˆ๊ฐ ์‹œํ•œ์„ ์•ž๋‘๊ณ  ์ด๋ฃจ์–ด์ง„ ๊ทน์ ์ธ ํ•ฉ์˜์˜€์Šต๋‹ˆ๋‹ค.

ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ ์ด๋ž€์ด 2์ฃผ๊ฐ„์˜ ํœด์ „ ๊ธฐ๊ฐ„ ๋™์•ˆ ํ˜ธ๋ฅด๋ฌด์ฆˆ ํ•ดํ˜‘(Strait of Hormuz)์„ ์™„์ „ํžˆ ์žฌ๊ฐœ๋ฐฉํ•  ๊ฒƒ์„ ์•ฝ์†ํ–ˆ๋‹ค๊ณ  ๋ฐํ˜”์Šต๋‹ˆ๋‹ค. ์ด ํ•ดํ˜‘์€ ์ „ ์„ธ๊ณ„ ์„์œ , ๋น„๋ฃŒ, ๋ฐ˜๋„์ฒด ์ƒ์‚ฐ์šฉ ํ—ฌ๋ฅจ ๋“ฑ ์ฃผ์š” ๋ฌผํ’ˆ์˜ ์ด๋™์„ ๋ง‰๋Š” ๋ณ‘๋ชฉ ํ˜„์ƒ์„ ์ผ์œผ์ผœ ์„ธ๊ณ„ ๊ฒฝ์ œ์— ๋ง‰๋Œ€ํ•œ ์˜ํ–ฅ์„ ๋ฏธ์น˜๊ณ  ์žˆ์—ˆ์Šต๋‹ˆ๋‹ค. ํœด์ „ ๋ฐœํ‘œ์™€ ๋™์‹œ์— ๋ฏธ๊ตญ์˜ ์ด๋ž€ ํญ๊ฒฉ์€ ๊ฑฐ์˜ ์ฆ‰์‹œ ์ค‘๋‹จ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

๋ฐ์ด๋น„๋“œ ์ƒ์–ด ๊ธฐ์ž๋Š” โ€œ์‹œ์žฅ์ด๋“  ๊ตฐ๋ถ€๋“  ๊ฑธํ”„ ์ง€์—ญ์˜ ๋งŽ์€ ์‚ฌ๋žŒ๋“ค์ด ์—„์ฒญ๋‚œ ์•ˆ๋„๊ฐ์„ ๋А๊ผˆ๋‹คโ€๊ณ  ๋‹น์‹œ ๋ถ„์œ„๊ธฐ๋ฅผ ์ „ํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ด ํœด์ „์ด ์–ผ๋งˆ๋‚˜ ์‹ค์งˆ์ ์ด๋ฉฐ, ๋ฌด์—‡์„ ์˜๋ฏธํ•˜๋Š”์ง€์— ๋Œ€ํ•œ ์˜๋ฌธ์€ ์—ฌ์ „ํžˆ ๋‚จ์•„์žˆ์Šต๋‹ˆ๋‹ค.

์—‡๊ฐˆ๋ฆฌ๋Š” ํ•ด์„: ํ˜ธ๋ฅด๋ฌด์ฆˆ ํ•ดํ˜‘ ํ†ต์ œ๊ถŒ ๋…ผ๋ž€

ํœด์ „ ํ•ฉ์˜ ์†Œ์‹์ด ์ „ํ•ด์กŒ์„ ๋•Œ, ๋ฏธ๊ตญ๊ณผ ์ด๋ž€์˜ ๋ฐœํ‘œ๋Š” ๋ฏธ๋ฌ˜ํ•œ ์ฐจ์ด๋ฅผ ๋ณด์˜€์Šต๋‹ˆ๋‹ค. ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์˜ ๋ฐœํ‘œ๋งŒ ๋†“๊ณ  ๋ณด๋ฉด ํ˜ธ๋ฅด๋ฌด์ฆˆ ํ•ดํ˜‘์ด ์ „์Ÿ ์ „์ฒ˜๋Ÿผ ์™„์ „ํžˆ ๊ฐœ๋ฐฉ๋  ๊ฒƒ์ฒ˜๋Ÿผ ๋“ค๋ ธ์Šต๋‹ˆ๋‹ค.

๊ทธ๋Ÿฌ๋‚˜ ๋ช‡ ๋ถ„ ๋’ค ์ด๋ž€ ์™ธ๋ฌด์žฅ๊ด€ ์•„๋ฐ”์Šค ์•„๋ผ๊ธฐ(Abbas Aragi)๊ฐ€ ๋ฐœํ‘œํ•œ ์„ฑ๋ช…์€ ๋‹ฌ๋ž์Šต๋‹ˆ๋‹ค. ๊ทธ๋Š” ์ด๋ž€์ด 2์ฃผ๊ฐ„ ๋ฐฉ์–ด ์ž‘์ „์„ ์ค‘๋‹จํ•  ๊ฒƒ์ด๋ฉฐ, ํ˜ธ๋ฅด๋ฌด์ฆˆ ํ•ดํ˜‘์„ ํ†ตํ•œ ์•ˆ์ „ ํ†ตํ–‰์€ ์ด๋ž€๊ตฐ๊ณผ์˜ ์กฐ์œจ์„ ํ†ตํ•ด์„œ๋งŒ ๊ฐ€๋Šฅํ•  ๊ฒƒ์ด๋ผ๊ณ  ๋ช…์‹œํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ์ด๋ž€์ด ์ง€๋‚œ 5~6์ฃผ๊ฐ„ ์žฅ์•…ํ–ˆ๋˜ ํ•ดํ˜‘ ํ†ต์ œ๊ถŒ์„ ๊ณ„์† ์œ ์ง€ํ•˜๊ฒ ๋‹ค๋Š” ์˜๋ฏธ๋กœ ํ•ด์„๋ฉ๋‹ˆ๋‹ค. ๋ฐ์ด๋น„๋“œ ์ƒ์–ด ๊ธฐ์ž๋Š” โ€œ์ด๋ž€์˜ ๋น„์ „ ์†์—์„œ๋Š” ์ด๋ž€๊ตฐ์ด ์—ฌ์ „ํžˆ ํ†ต์ œ๊ถŒ์„ ๋ณด์œ ํ•˜๋ฉฐ, ์–ด๋–ค ์†๋„๋กœ ์„ ๋ฐ• ํ†ต๊ณผ๋ฅผ ํ—ˆ์šฉํ• ์ง€, ํ†ตํ–‰๋ฃŒ๋ฅผ ๋ถ€๊ณผํ• ์ง€ ์•Œ ์ˆ˜ ์—†๋‹คโ€๊ณ  ์ง€์ ํ–ˆ์Šต๋‹ˆ๋‹ค.

ํ•œํŽธ, ์ด๋ฒˆ ์ „์Ÿ์˜ ์ฃผ์š” ์ฃผ์ฒด์˜€๋˜ ์ด์Šค๋ผ์—˜ ์—ญ์‹œ ๋ฐฑ์•…๊ด€์„ ํ†ตํ•ด ํ•ฉ์˜์— ๋™์˜ํ–ˆ๋‹ค๊ณ  ์ „ํ•ด์กŒ์Šต๋‹ˆ๋‹ค. ์ดํ›„ ์ด์Šค๋ผ์—˜ ์ด๋ฆฌ์‹ค์€ ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์˜ ๊ฒฐ์ •์„ ์ง€์ง€ํ•œ๋‹ค๋Š” ์„ฑ๋ช…์„ ๋ฐœํ‘œํ–ˆ์ง€๋งŒ, โ€œ์—ด๋ ฌํžˆ ํ™˜์˜ํ•˜๋Š” ๋ถ„์œ„๊ธฐ๋Š” ์•„๋‹ˆ์—ˆ๋‹คโ€๋Š” ํฅ๋ฏธ๋กœ์šด ํ‘œํ˜„์ด ์‚ฌ์šฉ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

์ผ์ด‰์ฆ‰๋ฐœ์˜ ์ˆœ๊ฐ„๋“ค: ์œ„๊ธฐ ๊ณ ์กฐ์˜ ์—ฐ๋Œ€๊ธฐ

์ด๋ฒˆ ํœด์ „์€ ๊ฐ‘์ž‘์Šค๋Ÿฝ๊ฒŒ ์ฐพ์•„์˜จ ๊ฒƒ์ด ์•„๋‹ˆ์—ˆ์Šต๋‹ˆ๋‹ค. ์ง€๋‚œ ๋ช‡ ์ฃผ๊ฐ„ ๋ฏธ๊ตญ๊ณผ ์ด๋ž€์€ ์ „์Ÿ์˜ ๋ฌธํ„ฑ๊นŒ์ง€ ๊ฐ€๋Š” ์•„์Šฌ์•„์Šฌํ•œ ๊ณผ์ •์„ ๊ฒช์—ˆ์Šต๋‹ˆ๋‹ค.

  1. ํ˜ธ๋ฅด๋ฌด์ฆˆ ํ•ดํ˜‘ ํ์‡„์™€ ํŠธ๋Ÿผํ”„์˜ ์ตœํ›„ํ†ต์ฒฉ: ํ˜ธ๋ฅด๋ฌด์ฆˆ ํ•ดํ˜‘ ๋ด‰์‡„๋กœ ์ธํ•œ ์ „ ์„ธ๊ณ„์  ๊ฒฝ์ œ ์••๋ฐ•์ด ๊ฐ€์ค‘๋˜์ž, ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ ์ด๋ž€์— ํ•ดํ˜‘์„ ๋‹ค์‹œ ์—ด๋ผ๋Š” 48์‹œ๊ฐ„ ์‹œํ•œ์„ ์ฃผ์—ˆ์Šต๋‹ˆ๋‹ค. ์ด๋ž€์€ ์ด๋ฅผ ๋ฌด์‹œํ–ˆ๊ณ , ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ 3์›” 26์ผ, โ€œํ˜‘์ƒ์„ ์‹œ์ž‘ํ•  10์ผ์˜ ์‹œ๊ฐ„์ด ์žˆ๋‹คโ€๋ฉฐ ๋˜ ๋‹ค๋ฅธ ์‹œํ•œ์„ ์ œ์‹œํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด๋ž€์€ ๊ณ„์†ํ•ด์„œ ๊ทธ๋ฅผ ๋ฌด์‹œํ•˜๊ณ  ์†Œ์…œ ๋ฏธ๋””์–ด์—์„œ ๋น„์›ƒ์—ˆ์Šต๋‹ˆ๋‹ค.

  2. F-15E ๊ฒฉ์ถ”์™€ ๊ทน์ ์ธ ๊ตฌ์กฐ ์ž‘์ „: ์ด ๋ชจ๋“  ๊ณผ์ •์€ ๊ธˆ์š”์ผ ๋ฐœ์ƒํ•œ F-15E ์ „ํˆฌ๊ธฐ ๊ฒฉ์ถ” ์‚ฌ๊ฑด์œผ๋กœ ์ž ์‹œ ์ค‘๋‹จ๋˜์—ˆ์Šต๋‹ˆ๋‹ค. ์กฐ์ข…์‚ฌ์™€ ๋ฌด๊ธฐ ๋‹ด๋‹น๊ด€์ด ํƒˆ์ถœํ–ˆ๊ณ , ๋ฌด๊ธฐ ๋‹ด๋‹น๊ด€์€ ์‹ค์ข…๋˜์—ˆ์Šต๋‹ˆ๋‹ค. ํ† ์š”์ผ, ์ด ์‹ค์ข…๋œ ๋ฌด๊ธฐ ๋‹ด๋‹น๊ด€์„ ๊ตฌ์ถœํ•˜๊ธฐ ์œ„ํ•œ โ€œ๋ฏฟ์„ ์ˆ˜ ์—†์„ ๋งŒํผ ๋Œ€๋‹ดํ•˜๊ณ  ๊ฐ’๋น„์‹ผโ€ ๊ตฌ์กฐ ์ž‘์ „์ด ์„ฑ๊ณต์ ์œผ๋กœ ์ด๋ฃจ์–ด์กŒ์Šต๋‹ˆ๋‹ค. ์ด๋Š” ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์—๊ฒŒ ์ž์‹ ๊ฐ์„ ๋ถˆ์–ด๋„ฃ์—ˆ๊ณ , ๊ทธ๋Š” ์ด๋ž€๊ณผ์˜ ํ˜‘๋ฐ•๊ณผ ํ˜‘์ƒ์œผ๋กœ ๋‹ค์‹œ ๋Œ์•„๊ฐ”์Šต๋‹ˆ๋‹ค.

  3. ํŠธ๋Ÿผํ”„์˜ ์œ„ํ˜‘ ์ˆ˜์œ„ ๊ฒฉ์ƒ: ์ „ํˆฌ๊ธฐ ๊ฒฉ์ถ” ์ „, ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ ์ด๋ž€์ด ๋ฌผ๋Ÿฌ์„œ์ง€ ์•Š์œผ๋ฉด ๋ฐœ์ „์†Œ์™€ ๋‹ค๋ฅธ ์ธํ”„๋ผ๋ฅผ ๊ณต๊ฒฉํ•˜๊ฒ ๋‹ค๊ณ  ์•”์‹œํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด๋ž€์€ โ€œ์šฐ๋ฆฌ์˜ ๋ฐœ์ „์†Œ๋ฅผ ํŒŒ๊ดดํ•˜๋ฉด ๊ฑธํ”„๋งŒ ์ฃผ๋ณ€์˜ ๋ฐœ์ „์†Œ์—๋„ ์ž‘๋ณ„์„ ๊ณ ํ•ด์•ผ ํ•  ๊ฒƒโ€์ด๋ผ๊ณ  ๋งž๋ฐ›์•„์ณค์Šต๋‹ˆ๋‹ค. ์‹ค์ข…๋œ ๊ณต๊ตฐ ๋Œ€๋ น์„ ๊ตฌ์ถœํ•œ ํ›„, ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ ๋‹ค์‹œ ์ž์‹ ์˜ ์š”๊ตฌ ์‚ฌํ•ญ์œผ๋กœ ๋Œ์•„์™”๊ณ , ์ผ๋ จ์˜ ์ถฉ๊ฒฉ์ ์ธ ์†Œ์…œ ๋ฏธ๋””์–ด ๊ฒŒ์‹œ๋ฌผ๋“ค์ด ์ด์–ด์กŒ์Šต๋‹ˆ๋‹ค.

โ€๋ฌธ๋ช… ์ „์ฒด๊ฐ€ ์ฃฝ์„ ๊ฒƒ์ด๋‹คโ€: ํŠธ๋Ÿผํ”„์˜ ๊ทน๋‹จ์  ์ˆ˜์‚ฌ

ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์˜ ์œ„ํ˜‘์€ ์ ์  ๋” ๊ตฌ์ฒด์ ์ด๊ณ  ๊ทน๋‹จ์ ์œผ๋กœ ๋ณ€ํ–ˆ์Šต๋‹ˆ๋‹ค.

  • ๋ถ€ํ™œ์ ˆ ์ผ์š”์ผ: ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ โ€œํ™”์š”์ผ์€ ์ด๋ž€์˜ ๋ฐœ์ „์†Œ์˜ ๋‚ ์ด์ž ๋‹ค๋ฆฌ์˜ ๋‚ ์ด ๋  ๊ฒƒ์ด๋‹ค. ๋ฏธ์นœ๋†ˆ๋“ค์•„, ๋‹น์žฅ ํ•ดํ˜‘์„ ์—ด์ง€ ์•Š์œผ๋ฉด ์ง€์˜ฅ์„ ๋ง›๋ณด๊ฒŒ ๋  ๊ฒƒ์ด๋‹ค. ๋‘๊ณ  ๋ด๋ผโ€๋ผ๋Š” ๊ฒŒ์‹œ๋ฌผ์„ ์˜ฌ๋ ธ์Šต๋‹ˆ๋‹ค. ์ƒ์–ด ๊ธฐ์ž๋Š” โ€œ์™ธ๊ต์ ์ด์ง€ ๋ชปํ•œ ์–ธ์–ด๋ฅผ ์ฐจ์น˜ํ•˜๊ณ ๋ผ๋„, ์ด๋Š” ๋Œ€ํ†ต๋ น์ด ์ด ์ „์Ÿ์— ํ›จ์”ฌ ๋” ๊นŠ์ด ๊ฐœ์ž…ํ•  ์˜์ง€๊ฐ€ ์žˆ์Œ์„ ์‹œ์‚ฌํ•˜๋Š” ๊ฒƒ์œผ๋กœ ๋ฐ›์•„๋“ค์—ฌ์กŒ๋‹คโ€๊ณ  ๋งํ–ˆ์Šต๋‹ˆ๋‹ค.

  • ์›”์š”์ผ: ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ ๋ฐฑ์•…๊ด€ ๋‚จ์ชฝ ์ž”๋””๋ฐญ์—์„œ ์—ด๋ฆฐ ๋ถ€ํ™œ์ ˆ ๋‹ฌ๊ฑ€ ๊ตด๋ฆฌ๊ธฐ ํ–‰์‚ฌ์—์„œ ์–ด๋ฆฐ์ด๋“ค๊ณผ ๋ถ€ํ™œ์ ˆ ํ† ๋ผ ์˜†์— ์„œ์„œ ์ด๋ž€์— ๋Œ€ํ•œ ์œ„ํ˜‘ ์ˆ˜์œ„๋ฅผ ๋”์šฑ ๋†’์˜€์Šต๋‹ˆ๋‹ค. ๊ทธ๋Š” โ€œ์ด๋ž€์˜ ๋‹ค๋ฆฌ, ๋ฐœ์ „์†Œ ๋“ฑ์„ ํŒŒ๊ดดํ•  ๊ฒƒ์ด๋ฉฐ, ๋ˆ„๊ตฌ๋„ ๋ณธ ์  ์—†๋Š” ๊ทœ๋ชจ๋กœ ๋ช‡ ์‹œ๊ฐ„ ์•ˆ์— ํ•ด๋‚ผ ๊ฒƒโ€์ด๋ผ๊ณ  ๋งํ–ˆ์Šต๋‹ˆ๋‹ค.

  • ํ™”์š”์ผ ์•„์นจ: ์˜ค์ „ 8์‹œ ์งํ›„, ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ ์ด์ „์˜ ๋ชจ๋“  ๋ฐœ์–ธ์„ ๋Šฅ๊ฐ€ํ•˜๋Š” ์ถฉ๊ฒฉ์ ์ธ ์†Œ์…œ ๋ฏธ๋””์–ด ๋ฉ”์‹œ์ง€๋ฅผ ์˜ฌ๋ ธ์Šต๋‹ˆ๋‹ค. โ€œ์˜ค๋Š˜ ๋ฐค ํ•œ ๋ฌธ๋ช… ์ „์ฒด๊ฐ€ ์‚ฌ๋ผ์งˆ ๊ฒƒ์ด๋ฉฐ, ๋‹ค์‹œ๋Š” ๋Œ์•„์˜ค์ง€ ๋ชปํ•  ๊ฒƒ์ด๋‹ค. ๋‚˜๋Š” ๊ทธ๋Ÿฐ ์ผ์ด ์ผ์–ด๋‚˜๊ธธ ์›์น˜ ์•Š์ง€๋งŒ, ์•„๋งˆ ๊ทธ๋ ‡๊ฒŒ ๋  ๊ฒƒ์ด๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ด์ œ ์šฐ๋ฆฌ๋Š” ๋‹ค๋ฅด๊ณ , ๋” ๋˜‘๋˜‘ํ•˜๊ณ , ๋œ ๊ธ‰์ง„์ ์ธ ์ •์‹ ์ด ์ง€๋ฐฐํ•˜๋Š” ์™„์ „ํ•˜๊ณ  ์ด์ฒด์ ์ธ ์ •๊ถŒ ๊ต์ฒด๋ฅผ ์ด๋ฃจ์—ˆ์œผ๋‹ˆ, ํ˜๋ช…์ ์œผ๋กœ ๋ฉ‹์ง„ ์ผ์ด ์ผ์–ด๋‚  ์ˆ˜๋„ ์žˆ๋‹ค. ๋ˆ„๊ฐ€ ์•Œ๊ฒ ๋Š”๊ฐ€? ์˜ค๋Š˜ ๋ฐค ์•Œ๊ฒŒ ๋  ๊ฒƒ์ด๋‹ค.โ€

๋ฐ์ด๋น„๋“œ ์ƒ์–ด ๊ธฐ์ž๋Š” ์ด ๋ฐœ์–ธ์— ๋Œ€ํ•ด โ€œ์ด๋Š” 9์ฒœ๋งŒ ๋ช…์˜ ์‚ฌ๋žŒ๋“ค์„ ์ „๋ฉธ์‹œํ‚ค๊ฒ ๋‹ค๋Š” ์œ„ํ˜‘์ด์—ˆ๋‹ค. ๋Œ€ํ†ต๋ น์€ ๋‹จ์ˆœํžˆ ์ด๋ž€ ์ •๋ถ€๋ฅผ ํŒจ๋ฐฐ์‹œํ‚ค๊ฑฐ๋‚˜ ํ•ต ์‹œ์„ค์„ ๊ณต๊ฒฉํ•˜๋Š” ๊ฒƒ์„ ๋„˜์–ด, ๋ณธ์งˆ์ ์œผ๋กœ ๊ฑฐ๋Œ€ํ•œ ์ „์Ÿ ๋ฒ”์ฃ„๋ฅผ ์ €์ง€๋ฅด๊ฒ ๋‹ค๋Š” ์ด์•ผ๊ธฐ๋ฅผ ํ•˜๊ณ  ์žˆ์—ˆ๋‹คโ€๊ณ  ๋ถ„์„ํ–ˆ์Šต๋‹ˆ๋‹ค.

๊ด‘๋ฒ”์œ„ํ•œ ๋น„ํŒ๊ณผ ๊ฒฝ๊ณ : โ€œ๋Œ€ํ†ต๋ น์€ ์ œ์ •์‹ ์ด ์•„๋‹ˆ๋‹คโ€

ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์˜ ๊ทน๋‹จ์ ์ธ ๋ฐœ์–ธ์€ ์ •์น˜์  ์ŠคํŽ™ํŠธ๋Ÿผ ์ „๋ฐ˜์— ๊ฑธ์ณ ๊ด‘๋ฒ”์œ„ํ•œ ๋น„ํŒ๊ณผ ์šฐ๋ ค๋ฅผ ๋ถˆ๋Ÿฌ์ผ์œผ์ผฐ์Šต๋‹ˆ๋‹ค.

  • ๋ฏผ์ฃผ๋‹น: ๋งŽ์€ ๋ฏผ์ฃผ๋‹น ์˜์›๋“ค์€ โ€œํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ ์ œ์ •์‹ ์ด ์•„๋‹ˆ๋‹ค,โ€ โ€œ๋Œ€ํ†ต๋ น์ง์„ ์ˆ˜ํ–‰ํ•  ์ž๊ฒฉ์ด ์—†๋‹คโ€๋ฉฐ ๋‚ด๊ฐ์ด ๋Œ€ํ†ต๋ น์˜ ์ง๋ฌด ์ˆ˜ํ–‰ ๋ถˆ๋Šฅ์„ ํŒ๋‹จํ•˜์—ฌ ๋ถ€ํ†ต๋ น์—๊ฒŒ ์ž„์‹œ ๊ถŒํ•œ์„ ๋ถ€์—ฌํ•˜๋Š” ์ˆ˜์ •ํ—Œ๋ฒ• 25์กฐ(25th Amendment) ๋ฐœ๋™์„ ์ด‰๊ตฌํ–ˆ์Šต๋‹ˆ๋‹ค.

  • MAGA ์šด๋™ ๋‚ด๋ถ€์˜ ๋น„ํŒ: ์‹ฌ์ง€์–ด ํŠธ๋Ÿผํ”„ ์ง€์ง€์ธต ๋‚ด๋ถ€์—์„œ๋„ ๋น„ํŒ์˜ ๋ชฉ์†Œ๋ฆฌ๊ฐ€ ๋‚˜์™”์Šต๋‹ˆ๋‹ค. ๊ทน์šฐ ๋…ผ๊ฐ ์•Œ๋ ‰์Šค ์กด์Šค(Alex Jones)๋Š” โ€œ์ˆ˜์ •ํ—Œ๋ฒ• 25์กฐ ๋ฐœ๋™์€ ํƒ„ํ•ต๋ณด๋‹ค ์–ด๋ ต๋‹คโ€๋ฉด์„œ๋„ โ€œํŠธ๋Ÿผํ”„๊ฐ€ ๊ฑด๊ฐ• ๋ฌธ์ œ๋ฅผ ๊ฒช๊ณ  ์žˆ๋‹ค๊ณ  ๊ณต๊ฐœ์ ์œผ๋กœ ๋ณด๊ณ ํ•˜๊ณ  ํŽœ์Šค ๋ถ€ํ†ต๋ น์ด ๋Œ€์‹ ํ•ด์•ผ ํ•œ๋‹ค. ๋ฌธ์ž ๊ทธ๋Œ€๋กœ ๊ทธ๋Ÿฐ ์ผ์ด ํ•„์š”ํ•  ์ •๋„๋กœ ์‹ฌ๊ฐํ•˜๋‹คโ€๊ณ  ๋งํ–ˆ์Šต๋‹ˆ๋‹ค. ๋งˆ์กฐ๋ฆฌ ํ…Œ์ผ๋Ÿฌ ๊ทธ๋ฆฐ(Marjorie Taylor Green)๊ณผ ์บ”๋””์Šค ์˜ค์›ฌ์Šค(Candace Owens) ๊ฐ™์€ ์ธ๋ฌผ๋“ค๋„ ์ˆ˜์ •ํ—Œ๋ฒ• 25์กฐ ๋ฐœ๋™์„ ์–ธ๊ธ‰ํ•˜๋ฉฐ ํŠธ๋Ÿผํ”„๋ฅผ โ€œ๋Œ€๋Ÿ‰ ํ•™์‚ด์ ์ธ ๋ฏธ์น˜๊ด‘์ด(genocidal lunatic)โ€œ๋ผ๊ณ  ๋น„๋‚œํ–ˆ์Šต๋‹ˆ๋‹ค.

  • ํ„ฐ์ปค ์นผ์Šจ๊ณผ ๋ก  ์กด์Šจ ์ƒ์›์˜์›: ๋ณด์ˆ˜ ๋…ผ๊ฐ ํ„ฐ์ปค ์นผ์Šจ(Tucker Carlson)์€ ํŠธ๋Ÿผํ”„์˜ ๊ฒŒ์‹œ๋ฌผ์„ โ€œ๋ชจ๋“  ๋ฉด์—์„œ ๋น„์—ดํ•˜๋‹ค(vile)โ€œ๊ณ  ๋ฌ˜์‚ฌํ•˜๋ฉฐ, โ€œ๋Œ€ํ†ต๋ น๊ณผ ์ง์ ‘ ์ ‘์ด‰ํ•˜๋Š” ์‚ฌ๋žŒ๋“ค์€ โ€˜๋‚˜๋Š” ์‚ฌ์ž„ํ•˜๊ฒ ๋‹คโ€™, โ€˜์ด๋Ÿฐ ๋ช…๋ น์„ ๋”ฐ๋ฅด์ง€ ์•Š์„ ๊ฒƒ์ด๋‹คโ€™, โ€˜์ด๊ฒƒ์„ ๋ง‰๊ธฐ ์œ„ํ•ด ๋ฒ•์ ์œผ๋กœ ํ•  ์ˆ˜ ์žˆ๋Š” ๋ชจ๋“  ๊ฒƒ์„ ํ•  ๊ฒƒ์ด๋‹ค. ์™œ๋ƒํ•˜๋ฉด ์ด๊ฒƒ์€ ๋ฏธ์นœ ์ง“์ด๊ณ , ๋ช…๋ น์ด ๋‚ด๋ ค์ง€๋ฉด ๋‚˜๋Š” ์‹คํ–‰ํ•˜์ง€ ์•Š์„ ๊ฒƒ์ด๋‹คโ€™๋ผ๊ณ  ๋งํ•ด์•ผ ํ•œ๋‹คโ€๊ณ  ์ฃผ์žฅํ–ˆ์Šต๋‹ˆ๋‹ค. ๋น„๊ต์  ๋Œ€ํ†ต๋ น ๋น„ํŒ์— ์‹ ์ค‘ํ–ˆ๋˜ ๋ก  ์กด์Šจ(Ron Johnson) ์œ„์Šค์ฝ˜์‹  ์ƒ์›์˜์›์กฐ์ฐจ โ€œ์ด๊ฒƒ์€ ๋ฏธ๊ตญ์ธ์ด ๋งํ•˜๋Š” ๋ฐฉ์‹์ด ์•„๋‹ˆ๋ฉฐ, ๋Œ€ํ†ต๋ น์ด ๋งํ•ด์•ผ ํ•˜๋Š” ๋ฐฉ์‹๋„ ์•„๋‹ˆ๋‹ค. ์šฐ๋ฆฌ๋Š” ์ด๋ž€ ๊ตญ๋ฏผ๊ณผ ์ „์Ÿ ์ค‘์ด ์•„๋‹ˆ๋‹ค. ์šฐ๋ฆฌ๋Š” ๊ทธ๋“ค์„ ํ•ด๋ฐฉ์‹œํ‚ค๋ ค ๋…ธ๋ ฅํ•˜๊ณ  ์žˆ๋‹คโ€๊ณ  ๋น„ํŒํ–ˆ์Šต๋‹ˆ๋‹ค.

์ทจ์•ฝํ•œ ํ‰ํ™”: ๋๋‚˜์ง€ ์•Š์€ ์ˆ™์ œ๋“ค

์ด๋ฒˆ ํœด์ „์€ ๋‹น์žฅ์˜ ์ „๋ฉด์ „ ์œ„๊ธฐ๋ฅผ ๋ชจ๋ฉดํ•˜๊ฒŒ ํ•ด์ฃผ์—ˆ์ง€๋งŒ, ๊ทธ ์‹ค์งˆ์ ์ธ ์˜๋ฏธ์™€ ์ง€์† ๊ฐ€๋Šฅ์„ฑ์€ ์—ฌ์ „ํžˆ ๋ถˆํ™•์‹คํ•ฉ๋‹ˆ๋‹ค.

  • ํœด์ „์˜ ์ทจ์•ฝ์„ฑ: ๋ฐ์ด๋น„๋“œ ์ƒ์–ด ๊ธฐ์ž๋Š” โ€œํœด์ „์€ ๋ณธ์งˆ์ ์œผ๋กœ ์ทจ์•ฝํ•˜๋‹ค. ๋ˆ„๊ฐ€ ์ด๋ฅผ ์œ„๋ฐ˜ํ• ์ง€ ์ง€์ผœ๋ด์•ผ ํ•œ๋‹คโ€๊ณ  ๊ฒฝ๊ณ ํ–ˆ์Šต๋‹ˆ๋‹ค. ๋˜ํ•œ ํ˜ธ๋ฅด๋ฌด์ฆˆ ํ•ดํ˜‘์˜ ์„ ๋ฐ• ํ†ตํ–‰๋Ÿ‰์ด ์‹ธ์›€์ด ๋ฐœ๋ฐœํ•˜๊ธฐ ์ „์ธ 2์›” 28์ผ ์ด์ „ ์ˆ˜์ค€์œผ๋กœ ํšŒ๋ณต๋ ์ง€ ์—ฌ๋ถ€๋„ ์ค‘์š”ํ•œ ์‹œํ—˜๋Œ€๊ฐ€ ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค.

  • ์ด๋ž€์˜ ์ง€๋ ›๋Œ€: ์ด๋ž€๊ตฐ์ด ์—ฌ์ „ํžˆ ํ•ดํ˜‘ ํ†ต์ œ๊ถŒ์„ ์ฅ๊ณ  ์žˆ๋‹ค๋Š” ์‚ฌ์‹ค์€ ์ค‘์š”ํ•œ ์˜๋ฏธ๋ฅผ ๊ฐ€์ง‘๋‹ˆ๋‹ค. ์ด๋Š” ์ด๋ž€์ด ์„ธ๊ณ„ ๋ฌด์—ญ์— ๋Œ€ํ•œ ๊ฐ•๋ ฅํ•œ ์ง€๋ ›๋Œ€(leverage)๋ฅผ ๋ฐœ๊ฒฌํ–ˆ์œผ๋ฉฐ, ์ด๋ฅผ ์‰ฝ๊ฒŒ ํฌ๊ธฐํ•˜์ง€ ์•Š์„ ๊ฒƒ์ž„์„ ์‹œ์‚ฌํ•ฉ๋‹ˆ๋‹ค.

  • ํ•ต ๋ฌธ์ œ์˜ ๋ฏธํ•ด๊ฒฐ: ํœด์ „์€ ๋” ํฐ ๋ฌธ์ œ์— ๋Œ€ํ•œ ํ˜‘์ƒ์„ ์œ„ํ•œ ์‹œ๊ฐ„๊ณผ ๊ณต๊ฐ„์„ ์ œ๊ณตํ•˜๊ธฐ ์œ„ํ•œ ์ˆ˜๋‹จ์ž…๋‹ˆ๋‹ค. ๊ฐ€์žฅ ์‹œ๊ธ‰ํ•œ ๋ฌธ์ œ๋Š” ์ด๋ž€์˜ ํ•ต ํ”„๋กœ๊ทธ๋žจ์ž…๋‹ˆ๋‹ค. ํŠธ๋Ÿผํ”„ ๋Œ€ํ†ต๋ น์€ ์ „์Ÿ ๋ฐœ๋ฐœ ์ „ โ€œ์ด๋ž€์ด ํ•ต ๋ฌผ์งˆ์„ ๋ณด์œ ํ•˜๋Š” ๊ฒƒ์„ ์šฉ๋‚ฉํ•  ์ˆ˜ ์—†๋‹คโ€๊ณ  ํ–ˆ๋‹ค๊ฐ€, โ€œ๋ณ„๋กœ ์‹ ๊ฒฝ ์“ฐ์ง€ ์•Š๋Š”๋‹ค. ์œ„์„ฑ์œผ๋กœ ๊ฐ์‹œํ•  ์ˆ˜ ์žˆ๋‹คโ€๊ณ  ๋ง์„ ๋ฐ”๊พธ๋Š” ๋“ฑ ์ผ๊ด€๋˜์ง€ ์•Š์€ ์ž…์žฅ์„ ๋ณด์—ฌ์™”์Šต๋‹ˆ๋‹ค. 2015๋…„ ์˜ค๋ฐ”๋งˆ ํ–‰์ •๋ถ€๊ฐ€ ์ด๋ž€๊ณผ ํ•ต ํ˜‘์ƒ์„ ํƒ€๊ฒฐํ•˜๋Š” ๋ฐ 2๋…„ ๋ฐ˜์ด ๊ฑธ๋ ธ๋‹ค๋Š” ์ ์„ ๊ฐ์•ˆํ•  ๋•Œ, ํ˜„์žฌ์˜ ์ „์Ÿ ์ƒํ™ฉ์—์„œ ๋” ํฐ ํ•ฉ์˜๋ฅผ ์ด๋Œ์–ด๋‚ด๋Š” ๊ฒƒ์€ ํ›จ์”ฌ ์–ด๋ ค์šธ ๊ฒƒ์ž…๋‹ˆ๋‹ค.

๋ฏธ๊ตญ์ด ์–ป์€ ๊ฒƒ๊ณผ ์žƒ์€ ๊ฒƒ: ๋๋‚˜์ง€ ์•Š์„ ์ „์Ÿ์˜ ์—ฌํŒŒ

์ด๋ฒˆ ์ „์Ÿ๊ณผ ํœด์ „์€ ๋ฏธ๊ตญ๊ณผ ์ด๋ž€, ๊ทธ๋ฆฌ๊ณ  ์ „ ์„ธ๊ณ„์— ๊นŠ์€ ์ƒ์ฒ˜์™€ ๊ตํ›ˆ์„ ๋‚จ๊ฒผ์Šต๋‹ˆ๋‹ค.

  • ๋ฏธ๊ตญ์˜ ์ œํ•œ์  ์„ฑ๊ณผ: ์ƒ์–ด ๊ธฐ์ž๋Š” โ€œ๋ฏธ๊ตญ๊ณผ ์ด์Šค๋ผ์—˜์€ ์ „์Ÿ ์ดˆ๊ธฐ์— ์ตœ๊ณ  ์ง€๋„์ž๋ฅผ ์ œ๊ฑฐํ•˜๊ณ , ๊ตฐ์‚ฌ, ์ •๋ณด, ํ•ต ์‹œ์„ค์„ ํ™ฉํํ™”์‹œํ‚ค๋ฉฐ, ๋ฏธ์‚ฌ์ผ ํ”„๋กœ๊ทธ๋žจ์„ ์ง€์—ฐ์‹œํ‚ค๋Š” ๋“ฑ ๋งŽ์€ ๊ฒƒ์„ ๋‹ฌ์„ฑํ–ˆ๋‹คโ€๊ณ  ํ‰๊ฐ€ํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ๋งŒ์•ฝ ์ด๋ฒˆ ํœด์ „์ด ํ˜ธ๋ฅด๋ฌด์ฆˆ ํ•ดํ˜‘์˜ ํ†ตํ–‰ ๋ณต์› ์™ธ์— ์ด๋ž€์— ์•„๋ฌด๋Ÿฐ ๋ณ€ํ™”๋„ ๊ฐ€์ ธ์˜ค์ง€ ์•Š๋Š”๋‹ค๋ฉด, ๋ฐฑ์•…๊ด€์€ ์ฃผ์š” ๋ชฉํ‘œ๋ฅผ ๊ฑฐ์˜ ๋‹ฌ์„ฑํ•˜์ง€ ๋ชปํ•˜๊ฒŒ ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค. ์ด๋ž€์€ ์—ฌ์ „ํžˆ ๊ตฐ๋ถ€์™€ ์‹ ์ • ์—˜๋ฆฌํŠธ๋“ค์— ์˜ํ•ด ํ†ต์น˜๋˜๊ณ  ์žˆ์œผ๋ฉฐ, ํ•ต ๋ฌผ์งˆ์„ ๋ณด์œ ํ•˜๊ณ  ์žˆ๊ณ , ๋ฏธ์‚ฌ์ผ๊ณผ ๋“œ๋ก  ์žฌ๊ฑด ๋Šฅ๋ ฅ์„ ๊ฐ€์ง€๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ๋ฌด์—‡๋ณด๋‹ค ์ด๋ž€์€ โ€œ๋ฏธ๊ตญ๊ณผ ์ด์Šค๋ผ์—˜์— ๋งž์„œ ์ €ํ•ญํ•  ์ˆ˜ ์žˆ์—ˆ๋‹ค๋Š” ์ธ์‹์„ ๊ฐ–๊ฒŒ ๋˜์—ˆ๋‹คโ€๋Š” ์ ์ด ์ค‘์š”ํ•ฉ๋‹ˆ๋‹ค.

  • ์žฅ๊ธฐ์ ์ธ ์˜ํ–ฅ:

    • ์ด๋ž€์˜ ์ƒˆ๋กœ์šด ํž˜: ์ด๋ž€์€ ์„ธ๊ณ„ ๋ฌด์—ญ์— ๋Œ€ํ•œ ์˜ˆ์ƒ์น˜ ๋ชปํ•œ ๊ฐ•๋ ฅํ•œ ์˜ํ–ฅ๋ ฅ์„ ๋ฐœ๊ฒฌํ–ˆ์Šต๋‹ˆ๋‹ค.
    • ๊ฑธํ”„ ๊ตญ๊ฐ€๋“ค์˜ ์ทจ์•ฝ์„ฑ: ๊ฑธํ”„ ๊ตญ๊ฐ€๋“ค์€ ๋ฏธ์‚ฌ์ผ ๊ณต๊ฒฉ์— ๋Œ€ํ•œ ์ทจ์•ฝ์„ฑ์„ ๊นจ๋‹ฌ์•˜์Šต๋‹ˆ๋‹ค.
    • ์„ธ๊ณ„ ๊ฒฝ์ œ์˜ ๋ถˆ์•ˆ์ •์„ฑ: ๋˜ ๋‹ค๋ฅธ ๊ณต๊ธ‰๋ง ์ทจ์•ฝ์ ์ด ๋“œ๋Ÿฌ๋‚˜๋ฉด์„œ ์„ธ๊ณ„ ๊ฒฝ์ œ๋Š” ๋”์šฑ ์ทจ์•ฝํ•ด์กŒ์Šต๋‹ˆ๋‹ค.
    • ๋ฏธ๊ตญ์˜ ์ค‘๋™ ๊ฐœ์ž…: ๋ฏธ๊ตญ์€ ์ค‘๋™์—์„œ ๋ฒ—์–ด๋‚˜๊ฒ ๋‹ค๋Š” ๋ชฉํ‘œ์—๋„ ๋ถˆ๊ตฌํ•˜๊ณ , ๋‹ค์‹œ๊ธˆ ์ƒ์ƒ ์ด์ƒ์˜ ๊ทœ๋ชจ๋กœ ์ค‘๋™์— ๊ฐœ์ž…ํ•˜๊ฒŒ ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.
    • ๋ฏธ๊ตญ์˜ ๊ตญ์ œ์  ์œ„์ƒ ์‹ค์ถ”: ๊ฐ€์žฅ ์ค‘์š”ํ•˜๊ฒŒ๋Š”, โ€œ๋ฏธ๊ตญ์ด ๊ณผ๊ฑฐ์— ๊ทผ๋ณธ์ ์œผ๋กœ ์ž๋น„๋กœ์šด ์ดˆ๊ฐ•๋Œ€๊ตญ์ด์—ˆ๋‹ค๋Š” ํ‰ํŒ์„ ๋งŽ์ด ์žƒ์—ˆ๋‹คโ€๋Š” ์ ์ž…๋‹ˆ๋‹ค. ๋Œ€ํ†ต๋ น์ด โ€œ์•ฝํ•œ ์ •๊ถŒ์„ ์ „๋ณต์‹œํ‚ค๊ณ , 9์ฒœ2๋ฐฑ๋งŒ ๋ช…์˜ ์ƒ๋ช…์„ ์œ„ํ˜‘ํ•  ์ˆ˜ ์žˆ๋‹คโ€๊ณ  ์ƒ๊ฐํ–ˆ์œผ๋ฉฐ, ์ด๋Ÿฌํ•œ ์œ„ํ˜‘์ด ์„ธ๊ณ„์—์„œ ๊ฐ€์žฅ ๊ฐ•๋ ฅํ•œ ๊ตฐ๋Œ€๋ฅผ ์ง€ํœ˜ํ•˜๋Š” ๋Œ€ํ†ต๋ น์œผ๋กœ๋ถ€ํ„ฐ ๋‚˜์™”๋‹ค๋Š” ์ธ์‹์ด ํ™•์‚ฐ๋˜์—ˆ์Šต๋‹ˆ๋‹ค. ํœ˜๋ฐœ์œ  ๊ฐ€๊ฒฉ์€ ๋‚ด๋ ค๊ฐ€๊ณ  ์„์œ ๋Š” ๋‹ค์‹œ ํ๋ฅด๊ฒ ์ง€๋งŒ, โ€œ์„ธ๊ณ„๊ฐ€ ๋ฏธ๊ตญ์„ ๊ณผ๊ฑฐ์ฒ˜๋Ÿผ ๋‹ค์‹œ ๋ณด๊ฒŒ ๋ ์ง€๋Š” ์˜ค๋žœ ์‹œ๊ฐ„์ด ๊ฑธ๋ฆด ๊ฒƒโ€์ด๋ผ๋Š” ์•”์šธํ•œ ์ „๋ง์ด ์ œ์‹œ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

์—ํ•„๋กœ๊ทธ: ๋˜ ๋‹ค๋ฅธ ์ธ์งˆ ์„๋ฐฉ

ํ•œํŽธ, ์ด๋ฒˆ ํœด์ „ ์†Œ์‹๊ณผ ํ•จ๊ป˜ ์ด๋ž€๊ณผ ์—ฐ๊ณ„๋œ ์ด๋ผํฌ ๋ฏผ๋ณ‘๋Œ€์— ๋‚ฉ์น˜๋˜์—ˆ๋˜ ๋ฏธ๊ตญ์ธ ์–ธ๋ก ์ธ ์…ธ๋ฆฌ ์ผ€์ด๋””์Šจ(Shelley KDson)์ด ์ผ์ฃผ์ผ๊ฐ„์˜ ์–ต๋ฅ˜ ๋์— ์„๋ฐฉ๋˜์—ˆ๋‹ค๋Š” ์†Œ์‹๋„ ์ „ํ•ด์กŒ์Šต๋‹ˆ๋‹ค. ๊ทธ๋…€๋Š” ์–ต๋ฅ˜๋œ ๋ฏผ๋ณ‘๋Œ€์› ์—ฌ๋Ÿฌ ๋ช…์˜ ์„๋ฐฉ๊ณผ ๊ตํ™˜๋˜์–ด ํ’€๋ ค๋‚ฌ์œผ๋ฉฐ, ์ด๋Š” ๋ฏธ๊ตญ๊ณผ ์ด์Šค๋ผ์—˜์ด ์ด๋ž€๊ณผ์˜ ์ „์Ÿ์„ ์‹œ์ž‘ํ•œ ์ดํ›„ ์ด๋ž€์˜ ์ง€์›์„ ๋ฐ›๋Š” ๋ฏผ๋ณ‘๋Œ€๋“ค์ด ๋ฏธ๊ตญ ๋ชฉํ‘œ๋ฌผ์„ ๊ณต๊ฒฉํ•  ์ˆ˜ ์žˆ๋‹ค๋Š” ๋ถˆ์•ˆ๊ฐ์ด ๊ณ ์กฐ๋˜๋Š” ๊ฐ€์šด๋ฐ ๋ฐœ์ƒํ–ˆ์Šต๋‹ˆ๋‹ค.

์ด๋ฒˆ ํœด์ „์€ ์ค‘๋™์˜ ๋ถˆ์•ˆ์ •ํ•œ ํ‰ํ™”๊ฐ€ ์–ผ๋งˆ๋‚˜ ์œ„ํƒœ๋กœ์šด์ง€๋ฅผ ๋‹ค์‹œ ํ•œ๋ฒˆ ๋ณด์—ฌ์ฃผ์—ˆ์Šต๋‹ˆ๋‹ค. ๋‹น์žฅ์˜ ์œ„๊ธฐ๋Š” ๋„˜๊ฒผ์ง€๋งŒ, ๊ทผ๋ณธ์ ์ธ ๊ฐˆ๋“ฑ์€ ์—ฌ์ „ํžˆ ํ•ด๊ฒฐ๋˜์ง€ ์•Š์€ ์ฑ„ ๋ณต์žกํ•œ ๊ตญ์ œ ์ •์„ธ์˜ ๊ทธ๋ฆผ์ž ์†์—์„œ ๋‹ค์Œ ๊ตญ๋ฉด์„ ๊ธฐ๋‹ค๋ฆฌ๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.